framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,2048,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,3072,1.1516212887234158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,64,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,65536,0.1616026692920261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,32,0.016322667400042217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,7168,0.13325866063435873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,5120,0.015754666593339708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,768,0.051232887638939746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,5120,0.012207110722859701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,5120,0.009661333428488838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,32,0.010286221901575724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,64,0.037703110112084284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,6144,0.013614222407341003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,256,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,3584,0.012865777644846173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,4096,0.007703999678293864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,6144,0.023390221926901076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,512,0.0033395555284288195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,768,0.003572444534964032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,51200,5.778436448838975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,5120,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,5120,0.039292444785436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,51200,0.06725422541300456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,1024,0.09819910923639934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,5120,0.010747555229398938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,128,0.003906666818592283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,8192,0.11343822214338516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,5120,0.00942311104800966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,2048,0.007962666451931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,64,0.0064293332397937775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,65536,1.2783280478583443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,8192,0.00701333334048589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,256,0.015471999843915304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,2560,0.027200889256265428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,51200,0.013083555632167392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,128,0.018432888719770644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,2560,0.053658667537901134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,3072,0.08348088794284397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,1536,0.08190577560000949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,7168,0.01442399952146742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,5120,0.006709333509206772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,10240,0.035389333963394165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,7168,0.011981333295504252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,1024,0.08280355400509305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,5120,0.08585600058237712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,768,0.008700444466537898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,10240,0.0454693337281545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,51200,0.11902844905853271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,4096,0.46820889578925234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,2560,0.19233422809176973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,512,0.0037493333220481873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,12288,0.00629333323902554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,7168,0.059679110844930015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,128,0.00300799993177255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,256,0.013320888910028668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,768,0.02103999919361538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,1536,0.05173066589567396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,65536,0.5526897642347548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,1536,0.01274311128589842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,4096,0.16457422574361166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,512,0.02455288834042019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,5120,0.01921244462331136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,8192,0.1250444385740492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,16384,0.014208889669842191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,5120,0.023521777656343248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,6144,0.031996445523367986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,2560,0.006786666810512543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,12288,0.06989866495132446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,256,0.0053671110007498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,3584,0.050310220983293324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,16384,0.023217777411142986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,768,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,65536,0.2854959964752197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,4096,0.04006844427850511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,1024,0.004478222380081813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,8192,0.09592533111572266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,2048,0.008647999829716152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,5120,0.01018133345577452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,7168,0.05658400058746338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,2048,0.006568000134494569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,16384,0.04844355583190918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,51200,0.05236444539493985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,4096,0.020437333318922255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,10240,0.03733333283000522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,768,0.007711999946170383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,8192,0.04302755660480923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,3584,0.03216000066863166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,512,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,32,0.0031795555518733132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,12288,0.1567360030280219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,256,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,51200,0.31698934237162274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,3072,0.08515822225146824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,2560,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,128,0.004589333302444882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,6144,1.3367786407470703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,512,0.030631999174753826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,4096,0.013688888814714221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,12288,0.0425511101881663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,2048,0.00618488879667388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,3584,0.009463110731707679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,10240,0.14976799488067627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,128,0.03621422251065572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,4096,0.10882044500774807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,65536,0.09285600317849053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,4096,0.0915635559293959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,64,0.023967110448413428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,51200,3.284131579928928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,256,0.003779555360476176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,2560,0.0377075539694892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,1024,0.016700444949997794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,10240,0.032497776879204646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,3584,0.787068419986301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,512,0.007648000286685095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,4096,0.009667555491129557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,10240,0.015765332513385348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,1024,0.005060444275538127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,16384,0.007237333390447829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,65536,0.32885421646965873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,51200,0.06599377923541598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,256,0.025477333201302424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,12288,0.7917111184861926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,65536,12.557200113932291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,3072,0.02588444451491038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,512,0.013016000390052795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,12288,0.1514764494366116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,7168,0.03649866580963135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,64,0.024511999554104272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,768,0.005503111001518037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,51200,0.2988648944430881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,5120,0.09040355682373047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,5120,0.012524444195959302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,12288,0.15541156133015951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,65536,0.0953306688202752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,512,0.015440889530711703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,768,0.005087999833954705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,128,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,768,0.0036115555299652945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,16384,0.01421866648726993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,3584,0.007816889219813878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,2048,0.022296888960732356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,1024,0.14897067017025417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,2048,0.022201778160201177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,256,0.005752888818581899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,16384,0.1880791054831611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,2048,0.10448355807198419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,10240,0.08886400196287368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,3072,0.012111110819710625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,128,0.00444444434510337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,1024,0.01591377788119846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,10240,0.03502577874395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,2560,0.006024888820118374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,3584,0.012952889005343119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,64,0.0052577778697013855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,65536,0.08634489112430149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,64,0.010120888551076254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,51200,1.3526923921373155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,5120,0.01827022267712487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,256,0.004129777765936322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,12288,0.370344877243042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,4096,0.14622578356001112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,2048,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,4096,0.08226578103171454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,10240,0.026276444395383198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,8192,0.22952977816263834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,256,0.008045333127180735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,3584,0.010135110881593492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,128,0.005127999931573868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,2048,0.009356444080670675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,51200,0.05548977851867676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,1536,0.19512621561686197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,6144,0.010335110955768162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,768,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,3584,0.009815110928482479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,32,0.004933333231343163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,65536,22.223917643229168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,10240,0.013473778135246701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,64,0.05466577741834852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,2560,0.02959644463327196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,32,0.06300444735421075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,2048,0.020008888509538438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,1024,0.012169777519173093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,16384,4.524870130750868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,8192,0.2659342288970947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,8192,0.06984977589713202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,3584,0.010404444403118556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,4096,0.010703110860453712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,16384,0.01443555619981554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,4096,1.3547582626342773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,1024,0.012987555729018318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,3584,0.011791110866599612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,2560,0.00980711148844825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,1536,0.006269333263238271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,768,0.035467555125554405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,768,0.004501333253251182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,2048,0.006711111300521427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,6144,0.021577777134047613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,51200,0.018561777141359117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,4096,0.440920008553399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,3072,0.25851554340786403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,512,0.004357333398527569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,1024,0.020039111375808716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,128,0.0155395550860299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,2048,0.0800666676627265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,1024,0.006296000132958095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,32,0.01797155539194743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,512,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,128,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,64,0.002712888850106133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,128,0.7811875343322754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,64,0.003980444537268745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,256,0.02929688824547662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,12288,0.09924799866146511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,512,0.008457777400811514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,3584,0.0351839992735121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,2048,0.005432888865470886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,768,0.016359110673268635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,1536,0.009411555197503831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,12288,0.04218577676349216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,3584,0.023195554812749226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,3584,0.011209777659840055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,7168,0.19352355268266466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,2560,0.041259557008743286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,16384,0.044534222947226636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,3072,0.022261333134439256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,5120,0.3168284363216824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,8192,0.927833769056532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,5120,0.20576355192396376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,2560,0.9410471386379666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,256,0.00702311098575592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,512,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,128,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,1024,0.028470221492979262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,768,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,8192,0.030893332428402368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,51200,0.1367866728040907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,1024,0.02240088913175795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,128,0.01606399979856279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,512,0.007146666447321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,7168,0.023984889189402264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,3584,0.20512799421946207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,768,0.01830577850341797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,256,0.02488533324665493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,256,0.005068444543414646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,768,0.012247111234400006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,3072,0.02830666634771559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,7168,0.020187555087937247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,7168,0.40579822328355575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,5120,0.016433777080641854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,3072,0.008583111067612966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,2048,0.00797333319981893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,7168,0.02853333287768894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,1024,0.007153777612580194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,1536,0.05617244376076592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,512,0.016028443972269695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,51200,0.3136800130208333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,512,0.004980444494220945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,128,0.010327111515733931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,51200,0.2154008944829305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,51200,0.5326044294569228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,4096,0.009137777818573846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,8192,0.03390489021937052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,128,0.025753777888086107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,3584,0.030714667505688135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,2560,0.1460497776667277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,4096,0.10405155685212876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,16384,0.04583644535806444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,256,0.00462755560874939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,4096,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,12288,0.018333334061834548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,3584,0.0105253333846728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,256,0.051402668158213295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,3072,0.010215111076831818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,256,0.028837333122889202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,5120,2.1780897776285806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,32,0.03720888826582167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,51200,8.251853942871094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,1024,0.009586667021115622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,3584,0.017995556195576985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,51200,1.6382675170898438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,3072,0.03149155444569058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,16384,0.9351973003811307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,3072,0.37050578329298234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,4096,0.027079110344250996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,3072,0.05048266715473599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,768,0.005580444302823808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,32,0.005436444448100195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,8192,0.03456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,51200,0.33230845133463544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,1536,0.004783111313978831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,1024,0.0069226668112807805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,8192,0.03423022230466207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,2560,0.08611822128295898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,51200,0.6519760025872124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,1024,0.04666755596796671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,3072,0.7628880076938205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,7168,0.03190577692455716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,7168,0.19453067249721953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,3584,0.010008888940016428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,12288,0.028378665447235107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,12288,0.4207022190093994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,128,0.008650666309727563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,128,0.007040888898902469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,10240,0.047394666406843394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,4096,0.014525334040323893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,2048,0.008091555701361762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,51200,0.9646355311075846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,6144,0.2659075525071886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,128,0.006797333144479328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,8192,0.23575022485521105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,64,0.09337777561611599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,64,0.011049778097205691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,512,0.003616888903909259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,64,0.002564444517095884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,7168,0.4733155568440755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,64,0.005158222383923001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,2560,0.010664000279373594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,10240,0.03143289022975498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,2560,4.809715694851346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,12288,0.028187556399239436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,4096,0.023219555616378784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,10240,0.06819199853473239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,10240,0.06693244642681546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,4096,0.010958222051461538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,8192,0.5935146543714735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,3072,0.008030222521887885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,65536,0.4640195634629991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,10240,0.3279342121548123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,7168,1.2830720477634006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,256,0.23953689469231498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,6144,0.22078843911488852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,2048,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,6144,0.046689778566360474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,3072,0.1401831176545885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,3072,0.008632000121805403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,32,0.008634666601816813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,768,0.042408890194363065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,4096,0.0475466681851281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,3072,0.008411555654472774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,65536,2.2825510236952042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,5120,0.07372444205813937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,7168,0.02129777769247691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,6144,0.159788449605306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,2560,0.04257244533962674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,1536,0.007248000138335758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,2048,0.006504888749784893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,8192,0.014357333381970724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,3072,0.009479111267460717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,12288,2.516934288872613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,8192,0.05286666750907898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,12288,0.012450666891203986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,128,0.007536888950400882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,5120,0.013960000541475085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,256,0.19923110802968344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,10240,0.12399910555945502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,2048,0.006075555665625467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,32,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,128,0.004176889028814104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,1024,0.0037919998996787598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,4096,0.05117155445946587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,8192,0.04125510983996921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,51200,0.2177635563744439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,768,0.003988444391224119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,3584,0.006680888848172293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,3072,0.012810666528013019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,5120,0.07803822225994535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,65536,0.1325182252460056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,51200,0.06984621948666044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,2048,0.007672000262472365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,4096,0.10650133424335056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,1024,0.026736888620588515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,3072,0.010568888651000129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,768,0.5697484546237522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,256,0.02972266740269131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,64,0.006029333505365584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,4096,0.008283555507659912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,2560,0.008060444560315874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,4096,0.3290719985961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,1536,0.009319999979601966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,65536,0.12840267022450766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,256,0.004201777693298128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,12288,0.3545955551995172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,4096,0.033703111939960055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,16384,0.0251217782497406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,16384,0.10519822438557942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,256,0.013750221994188098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,2048,0.06711555851830377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,256,0.0269777774810791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,6144,0.02163200080394745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,12288,0.016780444317393832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,7168,1.7383173836602106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,3072,0.02731288969516754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,64,0.0034702221552530923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,12288,0.11066666576597427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,7168,0.14067644543117946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,8192,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,64,0.005610666755172942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,8192,0.01498133275243971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,1536,0.058893336190117725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,7168,0.011965333587593503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,256,0.00490133340160052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,7168,0.11752888891432021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,12288,0.0383200016286638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,2048,0.006455111006895701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,16384,0.07561422056621976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,3584,0.6043039957682291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,2560,0.11668445004357232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,256,0.009882666998439366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,6144,0.01721955504682329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,2560,0.03297066688537598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,32,0.002777777819169892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,2048,0.015523556205961438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,65536,2.691263198852539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,12288,0.08345777458614773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,128,0.0035084444615576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,3584,0.011352889239788055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,7168,0.0184879998366038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,16384,0.1366951068242391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,12288,0.20160444577534994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,1024,0.11226844787597656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,7168,0.10521422492133246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,5120,0.009186666872766282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,7168,0.14084711339738634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,51200,1.4349626964992945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,3072,0.0962382223871019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,64,0.008547555241319869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,512,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,2560,0.050047109524408974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,768,0.006001777946949005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,3072,0.014601776997248331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,1024,0.08093421989017062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,32,0.007829333345095316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,128,0.007507555186748505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,3584,0.12203821871015762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,1536,0.006181333214044571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,1024,0.013023110727469126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,10240,3.4225635528564453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,51200,0.0655848913722568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,5120,0.030999110804663763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,64,0.012661333713266583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,10240,0.36329955524868435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,3584,0.050395554966396756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,8192,8.40554724799262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,65536,0.8421635627746582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,4096,0.02752799954679277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,7168,0.35935023095872665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,1024,0.006672888994216919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,3584,0.016460445192125108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,12288,0.006716444260544247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,12288,0.24900354279412162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,256,0.002883555574549569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,3072,0.007224000162548489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,2560,0.047376887665854565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,51200,0.3164844512939453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,1024,0.06386666827731662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,5120,0.1024542252222697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,2560,0.025361778007613286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,6144,3.6943511962890625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,65536,0.5141759978400336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,3584,0.011024889018800525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,256,0.021128000484572515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,1536,0.01885422236389584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,768,0.007977777885066139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,768,0.03378666771782769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,6144,0.009506666825877296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,256,0.009696889254781935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,1024,0.03298399845759074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,6144,0.03390222125583225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,32,0.0032266666077905228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,512,0.011161777708265515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,4096,0.19127289454142252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,2560,0.6599777539571127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,1536,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,3584,0.01809599995613098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,10240,0.18254044320848253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,2048,0.005445333404673471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,768,0.023410666320059035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,3072,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,8192,0.37035200330946183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,1536,0.05581600136227078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,32,0.009033777647548253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,7168,0.5557466612921821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,768,0.02366311185889774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,3584,0.12607555919223362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,256,0.003519111209445529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,768,0.01869066721863217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,512,0.0262773334980011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,2048,0.01640888883007897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,32,0.005701333284378052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,2560,0.053816000620524086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,128,0.01184266640080346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,51200,0.8893555535210503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,768,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,16384,0.7014257642957898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,3584,0.012245333029164208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,8192,0.03639644384384155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,6144,9.996088663736979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,512,0.010662222074137794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,65536,0.47347291310628253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,3584,0.03159644537501865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,2048,0.006743999818960826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,128,0.010433778166770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,51200,0.25572088029649526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,1536,0.023577777875794306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,51200,3.3322132958306203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,256,0.007252444823582967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,12288,0.023357333408461675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,256,0.015987555185953777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,51200,0.140754673216078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,64,0.002605333303411802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,64,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,512,0.005032000028424793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,512,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,2048,0.05508266554938423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,1024,0.04148355457517836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,2048,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,3072,0.08199466599358453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,8192,0.048159112532933555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,12288,0.04978222317165799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,4096,0.3438444402482774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,3584,0.01628444426589542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,32,0.025018667181332905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,768,0.008047999607192146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,2048,0.017257778180970084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,3072,0.012829333543777466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,512,0.006616888774765863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,5120,0.09499200185139973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,16384,0.030445333984163072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,8192,0.013819555441538492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,64,0.00564177789621883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,256,0.0063999998900625445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,64,0.006735999965005451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,10240,0.11886489391326904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,3072,0.015232000086042615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,3584,0.018561777141359117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,512,0.004863111095296012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,256,0.0033679999825027254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,65536,5.749995761447483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,12288,1.7360622617933485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,6144,0.030637333790461223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,2560,0.049599998527103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,16384,0.018572444717089336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,8192,0.13665244314405653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,3072,0.009659555223253038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,51200,0.06465778085920545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,2048,0.12796000639597574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,1536,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,12288,0.017326222525702584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,5120,0.09026044607162476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,4096,0.0724924471643236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,65536,0.3909946547614203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,8192,0.10962044530444676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,65536,0.5255022048950195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,2048,0.007975111405054728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,3072,0.0075457774930530125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,1536,0.006807111203670502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,4096,0.008227555288208855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,3072,0.060136000315348305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,1536,0.010455999937322406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,128,0.004942222187916438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,3072,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,7168,0.013017777767446307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,1024,0.030104888810051814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,12288,0.21519911289215088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,8192,0.019973332683245342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,51200,1.0924000210232205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,768,0.010313777459992303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,768,0.003765333443880081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,16384,0.12995999389224583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,2560,0.029059555795457628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,12288,0.044827553961012095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,16384,0.19753511746724448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,4096,0.06749155786302355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,65536,0.25576888190375435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,7168,0.06924266947640313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,256,0.026268444127506677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,16384,0.04077688852945963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,3072,0.06070311201943291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,2560,0.022589332527584497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,7168,0.1363111072116428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,51200,0.2191822263929579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,51200,0.05359289050102234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,2560,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,6144,0.03436000148455302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,10240,0.2615786658393012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,128,0.01976088848378923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,128,0.002895999906791581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,10240,0.04809955424732632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,8192,1.22980531056722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,6144,0.04623466730117798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,32,0.01510933372709486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,1536,0.030510223574108545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,51200,1.350431972079807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,10240,0.03355733222431607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,8192,0.010857777463065254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,128,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,51200,0.10798577467600505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,2560,0.01016088906261656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,12288,0.46085691452026367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,65536,0.348577790790134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,3072,0.024113777610990737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,64,0.007299555672539606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,7168,0.18203111489613852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,5120,0.2064675490061442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,12288,0.12202933099534775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,128,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,128,0.005070222334729301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,1536,0.029995554023318823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,4096,0.09303022093243069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,5120,0.021906667285495337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,128,0.0075839997993575195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,16384,0.09833244482676189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,65536,0.5455199877421061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,7168,2.1741981506347656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,3072,0.19030133883158365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,768,0.004466666529575984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,512,0.011384888655609555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,51200,0.053871999184290566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,64,0.005362666729423735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,8192,0.17309867011176217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,4096,0.07623466518190172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,8192,0.0510808891720242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,768,0.01936533384852939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,10240,0.013608888619475894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,1536,0.07146666447321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,4096,0.026912000444200303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,256,0.004937777916590373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,6144,0.30693422423468697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,5120,0.06241689125696818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,65536,0.25749511188930935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,8192,0.02085688875781165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,2560,0.009963555468453301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,128,0.002622222320901023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,256,0.01108888867828581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,2560,0.029339555237028334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,10240,0.04487911197874281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,1024,0.006082666830884085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,3072,0.010885333021481832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,3584,0.15709155135684544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,1536,0.05067111055056254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,8192,0.10150133238898383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,12288,0.018156444032986958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,4096,0.383424891365899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,3072,0.007557333343558841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,256,0.003640000190999773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,7168,1.416834619310167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,128,0.004104888687531154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,128,0.0032479998966058097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,6144,1.4034498002794054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,3072,0.02646844420168135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,5120,0.022630222969584998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,12288,0.02518400053183238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,16384,7.413342793782552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,2560,0.04151111178927951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,1536,0.048126220703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,3584,0.03018488817744785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,12288,0.17755733595954049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,64,0.005144888742102517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,1536,0.006616888774765863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,7168,0.45446046193440753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,8192,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,12288,0.12638844384087458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,7168,0.028392000330819026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,3584,0.015683554940753512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,12288,0.01830222209294637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,1024,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,12288,0.23412267367045084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,4096,0.06566666894488864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,51200,15.9491695827908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,3584,0.05337422092755636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,16384,0.03481066558096144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,1536,0.008798222574922774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,65536,0.07460088862313165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,32,0.007685333490371704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,7168,0.01905777719285753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,16384,0.023815999428431194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,5120,0.01792355544037289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,6144,0.019493333167499967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,2048,0.0051893335249688895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,8192,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,2048,0.7652293311225044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,65536,0.07525777816772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,7168,0.062385777632395424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,8192,0.16000266869862875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,4096,0.008461332983440822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,768,0.006189333481921091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,8192,0.21669511000315347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,2048,0.008763555851247575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,2048,0.007255111303594377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,65536,0.08660088645087348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,10240,3.8156941731770835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,12288,0.025007999605602686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,65536,2.113399929470486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,10240,0.04363288813167148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,4096,0.47339555952284074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,128,0.0035502223504914176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,6144,1.126736005147298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,51200,0.6646089024013943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,7168,0.03533688849873013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,3584,0.046816888782713145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,51200,0.047489778863059155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,128,0.012748444245921241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,512,0.008953777452309927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,512,0.0052888890107472735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,1536,0.010645333263609143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,5120,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,3584,0.006324444380071428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,16384,0.010191111101044549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,128,0.0026515554636716843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,16384,0.019001776973406475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,7168,0.3130142158932156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,3584,0.012505778008037143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,1024,0.003752888904677497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,512,0.03170844581392076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,51200,1.7255626254611547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,16384,0.449544005923801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,16384,0.009128000173303816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,65536,3.8648232354058156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,3584,0.02495999965402815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,8192,0.5919902059766982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,3584,0.1008346676826477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,12288,0.14645511574215359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,2560,0.00963288876745436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,6144,0.01144177797767851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,512,0.007119999991522894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,2048,0.013567999833159976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,10240,0.4992399745517307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,10240,0.4463671048482259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,3584,0.17698666784498426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,12288,0.05080355538262261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,7168,0.05752444267272949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,10240,0.03264266583654616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,3072,0.10771199729707505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,4096,0.01460533340771993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,51200,0.05429866578843859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,2048,0.006600889066855113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,5120,0.012015110916561551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,4096,0.008232889076073965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,128,0.006636444479227066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,2560,0.0158906661801868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,51200,0.06016621987024943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,16384,0.08351555797788833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,16384,0.06317866510815091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,32,0.0026542221506436667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,3072,0.014210666219393412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,256,0.01051111105415556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,256,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,2048,0.03703822361098395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,51200,3.007665846082899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,128,0.005450666778617435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,1024,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,256,0.009674666656388177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,2048,0.02091466718249851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,12288,0.017320000463061862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,1024,0.006680888848172293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,3584,0.0069422221018208405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,8192,0.1122773355907864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,10240,6.852732340494792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,4096,0.3244640032450358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,65536,0.9695911407470703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,512,0.009445333646403419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,2048,0.07721689012315539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,65536,1.4783297644721136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,65536,8.356910705566406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,1536,0.028379556205537584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,6144,0.013280889226330651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,12288,0.05419022176000807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,1024,0.006344888773229387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,10240,0.055870221720801465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,4096,0.19625511434343126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,64,0.03617066807217068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,2560,0.035629332065582275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,128,0.009744889206356471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,128,0.003039999968475766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,1024,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,1536,0.10058399703767563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,8192,0.04303466611438327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,4096,5.214839087592231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,2560,0.02423288921515147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,32,0.003452444449067116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,2048,0.012520888613329994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,64,0.004020444634887907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,128,0.0041706665522522396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,2560,0.008021333151393467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,4096,0.13311644395192465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,51200,0.2307493421766493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,1024,0.02477688921822442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,3584,2.3749894036187067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,1024,0.023822221491071913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,256,0.016139555308553908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,4096,0.0229120006163915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,65536,0.12235911687215169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,1024,0.005865777946180767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,5120,0.018569777409235638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,7168,0.01684888866212633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,768,0.03341955608791775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,4096,0.08059110906389025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,256,0.00871288859181934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,512,0.006525333142942852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,1024,0.021015110943052504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,2560,0.0070675553547011475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,1024,0.004295111116435793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,512,0.010861333045694562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,256,0.005201777650250329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,3072,0.13344000445471868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,4096,0.02478933334350586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,51200,17.710530598958332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,51200,1.134952015346951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,12288,0.049067556858062744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,2048,0.020289777053727042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,10240,0.039840890301598444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,2560,0.12782577673594156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,16384,0.008193777667151557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,1024,0.004277333203289243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,1024,0.015027556154463025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,8192,0.10159199767642552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,6144,0.17556711037953696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,3072,0.01811911165714264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,4096,0.23305511474609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,6144,0.09906578063964844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,7168,0.07568266656663683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,8192,0.06079466475380791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,64,0.003042666655447748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,8192,0.012438221938080259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,2560,0.05383733246061537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,128,0.005990222096443176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,5120,0.09074933661354913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,10240,0.06726044416427612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,3072,0.021771555145581562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,512,0.024193776978386775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,5120,1.419264899359809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,1536,0.13555467128753662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,512,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,512,0.003415999934077263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,768,0.1776533391740587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,2560,0.01128888875246048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,3072,0.01384888920519087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,12288,0.43084179030524355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,1536,0.9374835756089953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,10240,0.012616000241703458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,6144,1.8072516123453777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,7168,0.01721600029203627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,12288,0.10244977474212646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,3584,0.030741333961486816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,3584,0.06065599785910713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,3072,0.00811644477976693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,3072,0.011047110789351992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,6144,0.08095022042592366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,10240,0.018254222141371835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,65536,0.6802737977769641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,1536,0.02788177794880337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,256,0.006924444602595435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,5120,0.01439377831088172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,6144,0.6661671002705892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,12288,0.016383111476898193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,768,0.00406577769252989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,8192,0.11796532736884223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,1536,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,12288,0.22914134131537545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,5120,0.019517333971129525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,128,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,5120,0.19229066371917725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,5120,0.008791999684439765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,64,0.002970666728085942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,768,0.0034515555534097883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,65536,0.44689332114325625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,1024,0.08625955714119805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,2560,0.00740444411834081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,1536,0.006833777659469181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,51200,0.46911022398206925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,768,0.03125777840614319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,10240,0.0559084415435791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,12288,0.014119111829333834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,2560,0.00600355573826366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,1536,0.008384889198674096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,65536,4.1698426140679254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,8192,0.38757954703436953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,12288,2.5291466183132596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,64,0.0041946665280395085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,5120,0.1269351111518012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,12288,0.04355288876427544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,128,0.0026977778308921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,3072,0.011030221978823343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,4096,0.16573688719007704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,768,0.037350220812691584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,10240,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,64,0.0026444445053736367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,3072,0.0074417781498697065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,51200,0.10819733142852783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,256,0.005495999836259418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,16384,0.019052444232834708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,256,0.007999111380841997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,16384,0.7973813480801053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,10240,13.44325425889757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,1024,0.1869822210735745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,768,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,65536,1.4892515606350367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,2048,0.01204622205760744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,1536,0.020190222395790946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,12288,0.0716808901892768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,51200,0.04226488868395487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,51200,0.3403288788265652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,7168,0.012133333418104382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,10240,0.19408533308241102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,12288,0.03328711125585768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,51200,0.04524088899294535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,8192,0.03595111105177138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,64,0.00272444449365139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,8192,0.08673333459430271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,128,0.006146666490369373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,8192,0.06309955649905734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,8192,0.3028346697489421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,6144,0.020664888951513503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,32,0.0052977779673205475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,128,0.00980444418059455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,1536,0.004643555730581284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,12288,0.11485155423482259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,7168,0.05693511168162028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,768,0.003530666646030214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,256,0.023729777998394434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,8192,0.25682489077250165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,768,0.007647111184067196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,7168,0.012075555821259817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,5120,0.01444355481200748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,32,0.003080000066094928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,1536,0.024506666594081458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,128,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,256,0.008456000023418004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,32,0.002768888862596618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,3584,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,512,0.005968889014588461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,512,0.0033848887930313745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,5120,0.10842133230633205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,64,0.002593777866827117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,51200,0.03852533300717672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,1536,0.007888889147175683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,12288,0.08701422479417588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,512,0.14224710729387072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,128,0.002692444456948174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,51200,0.05793066819508871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,5120,0.1756524509853787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,4096,0.011352000137170156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,6144,0.034797334008746676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,3072,0.014485332700941296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,768,0.004351110921965705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,5120,0.10525511370764838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,1536,0.010215111076831818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,6144,1.7757742140028212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,51200,0.13771911462148032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,7168,0.17325866222381592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,512,0.020663999848895602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,6144,1.5428088506062825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,768,0.01608533329433865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,6144,0.018491556247075398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,51200,0.051613334152433604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,64,0.004567111117972268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,2048,0.023619555764728125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,3584,0.03602844476699829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,12288,1.6107021967569988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,512,0.008278222547637092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,6144,0.03016888764169481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,32,0.0063733334342638654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,32,0.0075937774446275495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,3584,0.012329777909649743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,65536,0.7479759852091471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,32,0.24212090174357095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,1024,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,1536,0.025136889682875738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,128,0.0031973332580592898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,5120,0.02974489000108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,1536,0.007149332927332983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,65536,0.4659084214104547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,2048,0.009115555220180089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,65536,1.2765430874294705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,3072,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,768,0.006337777607970768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,2560,0.04054044352637397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,12288,0.04660177893108792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,51200,0.4077777862548828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,64,0.002712888850106133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,256,0.0046471113132105935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,7168,0.151147551006741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,10240,0.3500444359249539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,51200,0.24644088745117188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,12288,0.13638222217559814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,4096,0.008795555267069075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,256,0.005530666559934616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,768,0.05974133147133721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,2560,0.009523555636405945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,2560,2.4010853237575955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,256,0.20865244335598418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,51200,0.0633208884133233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,2560,0.04061066773202684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,3584,0.010842666857772403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,65536,0.562172465854221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,256,0.045886221859190196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,6144,0.014498665928840637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,1024,0.0269804447889328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,2048,0.6424080000983344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,6144,0.006163555714819167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,768,0.0598506662580702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,2560,0.023966223001480103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,2048,0.008348444269763099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,32,0.002687999978661537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,65536,0.2560177909003364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,3072,0.3821449014875624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,6144,0.025349333882331848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,256,0.0053671110007498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,65536,0.06621422370274861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,256,0.015638222297032673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,128,0.012327999704413943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,2048,0.013861333330472311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,16384,0.7031475702921549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,768,0.15079200267791748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,256,0.0032915555768542816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,32,0.01642311116059621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,10240,0.2559777895609538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,64,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,3072,0.050831112596723765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,512,0.006711999989218182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,4096,0.011048888994587792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,6144,0.01203555530971951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,128,0.0027520000520679685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,1024,0.004951999833186467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,3584,0.010168000227875179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,12288,0.015032889114485847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,65536,1.7430204815334742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,3584,0.5706088807847765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,2048,0.03498844305674235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,5120,0.12550755341847739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,1024,0.021551999780866835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,3584,0.06824088758892484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,65536,0.05299733413590325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,128,0.004865777575307422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,8192,0.03583733240763346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,65536,0.06689777639177111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,12288,0.06930666499667697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,768,0.009568889107969072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,8192,0.01309866706530253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,5120,0.0162435554795795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,6144,0.027083555857340496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,10240,0.11493866973453098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,256,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,8192,0.03445511062939962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,3072,0.049132443136639066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,3072,0.2366488774617513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,65536,0.6877235306633843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,4096,0.06193511353598701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,768,0.030701332622104224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,2048,0.039592888620164655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,1024,0.005394666559166378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,32,0.009290667043791877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,1536,0.008384889198674096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,2560,0.025065778030289546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,3072,0.02200444373819563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,3072,0.08532800277074178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,2560,0.02030044462945726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,65536,0.2942088974846734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,2048,0.4130888779958089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,512,0.0035217776894569397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,2560,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,64,0.0028799999919202593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,3072,0.040511111418406166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,768,0.004048000193304486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,12288,0.020455110404226515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,10240,0.030087110069062974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,5120,0.009298666483826106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,51200,0.16323199537065294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,5120,0.2964968946244982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,2560,0.022082666556040447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,3072,0.010383110907342698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,512,0.004589333302444882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,1536,0.00638755535085996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,512,0.008591999610265097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,3584,0.009923555784755284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,1536,0.012461333639091916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,8192,0.01827111177974277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,768,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,5120,3.097229427761502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,10240,0.6297048992580837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,2560,0.007755555212497711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,3584,0.016912889149453905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,128,0.0034355556385384668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,2560,0.017473777135213215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,3072,1.0732933680216472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,2048,0.006057777752478917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,3584,0.04001155495643616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,1536,0.01091822236776352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,1024,0.04388799932267931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,3072,0.010093332992659675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,3584,0.009861333502663506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,1536,0.1359235578113132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,7168,0.2812008857727051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,5120,0.03510399990611606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,32,0.042318221595552236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,2560,0.03232889042960273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,768,0.006685333533419504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,768,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,7168,0.05680355760786268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,5120,0.012077333198653327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,4096,0.01370044466522005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,256,0.0047173334492577445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,512,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,3072,0.0109120003051228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,6144,0.05738133192062378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,128,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,16384,0.7592097918192545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,12288,0.31060889032151967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,1536,0.011798222031858234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,256,0.009677333136399588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,128,0.009322666459613377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,512,0.007002666592597961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,10240,2.515610588921441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,3584,0.018135999639829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,3072,0.010132444401582083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,51200,0.15779822402530247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,7168,0.035306665632459856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,5120,0.013764444324705334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,768,0.02913155489497715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,2560,0.10433333449893528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,7168,0.02363377809524536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,4096,0.03465066684616937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,65536,0.35482488738165957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,51200,0.7911555502149793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,1024,0.004606222112973531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,1536,0.00904799997806549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,32,0.008933333059151968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,5120,0.02696888811058468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,1536,0.005757333503829108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,12288,0.3170328934987386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,768,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,12288,0.03412977854410807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,7168,0.019256000717480976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,1536,0.03842488924662272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,1024,0.00922400007645289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,6144,0.012379555238617791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,2048,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,2560,1.0029164420233834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,768,0.2556568781534831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,256,0.0041955556306574075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,51200,0.13754133383433023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,5120,0.16975022686852348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,6144,0.021542222963439092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,256,0.041133334239323936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,10240,0.29066401057773167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,1536,0.007365333537260692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,10240,0.01371911085314221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,8192,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,10240,0.03356622325049506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,2048,0.011048888994587792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,7168,0.012672888735930124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,512,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,7168,0.1089751124382019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,1024,0.006039111150635614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,4096,0.017423111531469557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,512,0.003367111086845398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,2560,0.052536888255013355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,2048,0.010665777656767102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,256,0.3376106686062283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,6144,0.17954311105940077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,16384,5.485413445366754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,3072,0.24269954363505045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,64,0.0026622222115596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,10240,0.022082666556040447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,8192,0.013725333743625216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,2048,0.015433778365453085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,5120,0.01718844473361969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,512,0.13222932815551758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,32,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,51200,0.06300533480114408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,16384,0.012834666503800286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,10240,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,10240,0.2691360049777561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,1024,0.0125928885406918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,256,0.007575111256705389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,6144,0.011413333316644033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,3584,0.017831110292010836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,6144,0.30013155937194824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,65536,0.37087199423048234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,1536,0.023426666855812073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,256,0.020503110355801053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,256,0.0032755554550223877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,768,0.0063893331421746155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,12288,0.0255422227912479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,1536,0.0063733334342638654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,3072,0.009699555734793345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,2560,0.06001777781380547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,128,0.00314666661951277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,256,0.01920355525281694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,32,0.016401777664820354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,4096,0.014720888601409065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,512,0.004816888935036129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,128,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,6144,0.030082666211658057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,512,0.007335111498832703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,64,0.011123555401961008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,6144,0.027302222119437322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,64,0.003626666549179289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,5120,0.012028444144460889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,2048,0.016943110360039603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,65536,0.12063911226060654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,4096,0.0240639994541804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,4096,0.05520711011356778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,128,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,6144,0.18330132961273193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,4096,0.014869333969222175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,16384,0.2970782121022542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,8192,0.010185778141021729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,10240,0.021777777208222285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,8192,0.0320995549360911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,64,0.003125333330697484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,3584,0.22396532694498697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,12288,0.2302560011545817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,7168,0.12603378295898438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,1536,0.006626666833957036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,128,0.0032008888406885993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,3584,0.5072026782565647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,1536,0.14372266663445368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,16384,0.015538665983412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,32,0.003108444520168834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,6144,0.02149244480662876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,1536,0.01406222250726488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,768,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,51200,8.39713372124566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,5120,0.07499555746714275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,1536,0.022504000200165644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,5120,0.018495111001862418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,256,0.003240888938307762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,512,0.004863111095296012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,4096,0.012618666721714867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,5120,0.09309689203898112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,51200,0.04420266548792521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,128,0.007792000141408708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,3584,0.033210668298933245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,7168,0.010960000256697336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,2048,0.024642666180928547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,512,0.025750221477614507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,32,0.004875555634498596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,256,0.018564444449212816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,7168,0.01620533400111728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,1536,0.037808001041412354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,5120,0.03042311138576931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,8192,0.015065777632925244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,4096,0.012227555943859948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,51200,0.18114311165279814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,65536,0.07101688782374065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,768,0.014544889330863953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,2048,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,5120,0.2390657795800103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,10240,0.029818667305840388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,512,0.24481243557400176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,8192,0.7686435381571451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,768,0.007718222008811102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,256,0.04103466537263658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,64,0.005130666825506422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,1536,0.08790310886171128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,5120,0.011928888658682505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,1024,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,768,0.009724444813198512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,65536,1.27802488538954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,10240,0.04201422135035197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,8192,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,256,0.011696889168686338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,51200,0.3168391121758355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,4096,0.01404800017674764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,16384,0.02629511058330536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,51200,0.8456408712599012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,1536,0.16093956099616155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,3584,0.00870222184393141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,10240,0.019268444842762418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,12288,5.8408203125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,3584,0.10360444254345363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,8192,2.2113049825032554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,10240,0.021613332960340712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,51200,1.5736764272054036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,2048,0.023128888673252527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,128,0.024500444531440735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,32,0.027849776877297297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,32,0.003181333343187968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,65536,0.10815199878480698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,32,0.0032106666929192017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,16384,0.022871999277008906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,8192,0.037300444311565824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,32,0.005454222361246745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,3072,0.0069306666652361555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,1536,0.027628444963031348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,65536,0.8905217912462023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,256,0.005923555543025334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,6144,0.09402044614156087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,2048,0.008985777695973715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,1536,0.01274400038851632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,2048,0.01587199999226464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,5120,0.02725511127048069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,51200,0.06200177801979912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,3072,0.01738133364253574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,3584,0.17164711157480875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,8192,0.015770667129092746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,512,0.0036924444139003754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,6144,0.1827075613869561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,51200,0.21155733532375762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,1024,0.009553777674833933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,64,0.0236453331179089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,256,0.00332622230052948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,32,0.006694222076071634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,512,0.010063110954231685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,65536,4.121631198459201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,2560,0.1478417714436849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,65536,5.82515631781684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,6144,0.024558222956127588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,12288,0.05182844400405884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,5120,0.009816889133718278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,12288,0.030679110023710463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,3072,1.0270089043511283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,128,0.015271999769740634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,3072,0.012144000165992312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,1024,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,16384,0.03995910949177212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,3584,0.010948444406191507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,5120,0.6858266724480523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,65536,2.9607787662082248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,1024,0.017215111189418368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,1536,0.1750026676389906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,256,0.0032337777730491427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,3584,0.01569688816865285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,512,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,3584,0.051761776208877563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,4096,0.022508444057570562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,768,0.040863109959496394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,512,0.004980444494220945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,64,0.005006222261322869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,51200,6.340717315673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,4096,0.20412888791826037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,128,0.0033119999700122406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,64,0.002573333266708586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,8192,0.06811288992563884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,12288,5.339950985378689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,128,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,51200,0.28724267747667104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,3072,0.11676622761620416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,51200,4.552486419677734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,16384,0.07036978006362915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,6144,1.8102152082655165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,6144,0.01556799974706438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,5120,0.04888088835610283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,64,0.007291555404663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,4096,3.3834959665934243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,768,0.43552266226874453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,4096,0.027095110880004034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,768,0.004004444513056013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,7168,0.015217777755525378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,51200,2.4346834818522134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,512,0.05655022462209066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,128,0.0025600000388092464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,256,0.019871110717455547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,1536,0.01722133325205909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,2560,0.012082666986518435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,2560,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,7168,0.1299928824106852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,2560,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,7168,0.017481777403089736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,7168,0.03770222266515096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,4096,0.012224888636006249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,4096,0.0455022222465939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,3072,0.008548444343937768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,16384,3.3208141326904297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,65536,0.6369484265645345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,51200,0.391612450281779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,8192,0.04850577645831638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,8192,0.01627200014061398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,3584,0.12184800042046441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,32,0.04393600092993843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,6144,0.05009866754213969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,4096,0.011413333316644033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,5120,0.14954755041334364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,2048,0.04231111208597819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,2048,0.009462222456932068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,6144,0.01975555552376641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,2560,0.01520088811715444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,1024,0.011750222080283694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,1024,0.011369778050316704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,2560,0.022816888160175745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,256,0.0063484443558586975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,12288,0.09433422485987346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,5120,0.02160533269246419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,8192,0.012245333029164208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,6144,0.18690221839480928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,1536,0.026700443691677515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,2048,1.8546863132052953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,8192,0.2537048922644721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,4096,0.010238221950001186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,1536,0.026820444398456152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,256,0.012472888661755456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,3072,0.01923733287387424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,16384,0.10876799954308404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,1536,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,2048,0.2562062210506863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,768,0.006478222293986215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,4096,0.02787289023399353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,6144,0.017122666041056316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,128,0.0029031110720502008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,3072,0.009170666337013245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,65536,1.0770595338609483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,2048,0.008632000121805403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,512,0.011345778074529437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,4096,0.03635111120012071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,512,0.007365333537260692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,128,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,2560,0.007113777928882175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,2048,0.008002666963471307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,256,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,128,0.005277333160241445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,12288,0.5514764255947536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,768,0.011328000161382886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,65536,4.461929745144315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,5120,0.014504889647165934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,3072,0.0653466648525662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,4096,0.16137422455681696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,65536,0.11918755372365315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,256,0.037977778249316745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,51200,0.12422400050693089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,4096,0.01163822246922387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,1024,0.02137511140770382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,1536,0.1647626691394382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,65536,0.7729057735866971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,6144,0.022650667362742957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,4096,0.01330311099688212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,3584,0.007801777786678738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,128,0.015020444989204407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,32,0.010727110836240979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,4096,0.15247110525767008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,3584,0.011164444188276926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,10240,0.014450665977266101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,10240,0.10193689001931085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,768,0.006699555450015598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,3584,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,51200,5.4839964972601996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,128,0.007393777370452881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,10240,0.49159733454386395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,2560,0.04145066605673896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,2560,0.009341333475377824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,3584,0.011317333413494958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,5120,0.012321777641773224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,2048,0.005371555685997009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,65536,0.3793493376837836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,3584,0.009381333159075843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,5120,0.03551377852757772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,2560,0.011075555450386472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,10240,0.02608355548646715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,128,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,4096,0.026643556025293138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,6144,0.02424088948302799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,128,0.0064071110553211635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,6144,0.06891022125879924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,51200,11.64097425672743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,2560,0.007528000407748752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,8192,0.0365608897474077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,256,0.015807110402319167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,2048,0.016639111770523917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,512,0.01110311100880305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,128,0.007755555212497711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,51200,0.19820088810390898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,128,0.02533866630660163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,128,0.009015111459626092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,5120,0.018931556079122756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,1024,0.010687111152542962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,256,0.005129777722888523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,1536,0.00980444418059455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,1024,0.011791110866599612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,1024,0.030174222257402208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,1536,0.006017777654859755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,12288,0.13107732931772867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,6144,0.009729777773221334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,32,0.04904799991183811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,768,0.023312888211674158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,3072,0.059106667836507164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,51200,0.08408888843324448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,2560,0.07823022206624348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,16384,0.0730924473868476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,4096,0.010791999598344168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,64,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,2048,0.0138586668504609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,256,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,6144,0.0427706672085656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,3072,0.007782222496138678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,12288,0.8956613540649414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,256,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,6144,0.09170489178763495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,2048,0.02142400046189626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,7168,0.021313778228229944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,256,0.1920293304655287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,65536,1.741421381632487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,5120,0.01184800018866857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,12288,0.041949331760406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,12288,0.045087112320793994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,3584,0.7945289081997342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,8192,53.07329305013021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,8192,0.0306328899330563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,16384,0.16819111506144205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,2048,0.025959110922283594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,1536,0.008982222113344405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,65536,0.07436888747745089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,2048,0.010142222046852112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,1024,0.005852444304360284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,16384,17.294140285915798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,128,0.004613333278232151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,128,0.005195555587609609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,64,0.008370666868156856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,4096,0.019693333241674636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,5120,0.015740444262822468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,6144,0.10435644123289321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,128,0.13490845097435847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,8192,0.29166311687893337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,512,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,1024,0.021913778450753953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,128,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,12288,0.05009866754213969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,51200,0.03275733192761739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,1024,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,12288,0.513079113430447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,2048,0.009574222067991892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,65536,44.329325358072914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,8192,0.05160266823238797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,16384,0.007187555233637492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,5120,0.013218666944238873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,2048,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,12288,0.024546666277779475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,768,0.0240320000383589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,10240,0.8417413499620227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,256,0.003915555361244413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,128,0.044016887744267784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,51200,0.17618844244215223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,10240,0.02400355537732442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,65536,0.253532436158922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,32,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,4096,0.04184444414244758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,6144,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,8192,0.009826666778988307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,2560,0.01888799998495314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,16384,5.143668280707465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,1536,0.013006221916940479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,10240,0.04008977942996555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,51200,0.08463733063803779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,5120,0.010176888770527309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,2048,0.0052639999323421055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,1536,0.04193422198295593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,256,0.03659111261367798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,2560,0.10088355673684014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,512,0.003564444267087512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,32,0.003012444410059187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,6144,0.005902222047249476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,4096,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,1536,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,512,0.032799998919169106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,256,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,51200,0.19332888391282824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,12288,4.31330829196506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,51200,0.5402195718553331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,2048,0.010598222414652506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,512,0.004490666505363253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,768,0.005850666513045629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,10240,0.007407111426194509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,12288,0.007033777733643849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,7168,0.6735857857598199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,768,0.008018666671382057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,2560,0.03157066636615329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,128,0.006606222026877933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,5120,0.018168000711335074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,7168,0.013092444174819522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,2560,0.1295777824189928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,5120,0.015827554795477126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,2560,0.007198221981525421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,1536,0.11767555607689752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,10240,1.1181217829386394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,12288,0.04707555638419258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,768,0.06133777565426297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,6144,0.017530666457282174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,4096,0.04364533225695292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,512,0.005903111149867375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,5120,0.024269334144062463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,1024,0.00553955551650789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,16384,0.044554667340384595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,2048,0.007181333170996771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,2560,0.12991289297739664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,12288,0.20671733220418295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,5120,0.019313777486483257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,12288,0.18589422437879774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,10240,0.01370044466522005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,65536,0.4178044531080458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,65536,3.2748764885796438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,2560,0.009240888886981541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,5120,0.025832888152864244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,7168,0.03384533193376329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,128,0.0059573331640826324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,7168,0.13948267036014134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,2560,0.016723554995324876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,1536,0.4155617819892035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,3072,0.01052977806992001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,3584,0.01867999964290195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,2048,0.09956266482671101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,4096,0.016287111573749118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,5120,0.1288506719801161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,5120,0.009217778013812171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,12288,0.03679910964436001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,12288,0.18012355433570015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,65536,0.10412889056735569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,8192,0.014841778410805596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,3584,0.027540443672074214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,16384,0.041357331805759005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,51200,0.1737040016386244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,65536,0.039288888374964394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,1536,0.04228533307711283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,3584,0.024588444166713293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,32,0.0031306667046414483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,3584,0.04426755507787069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,16384,0.31653422779507107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,768,0.08631200260586208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,64,0.13535911507076687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,16384,0.5643226835462782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,64,0.02128177715672387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,4096,0.009015999734401703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,1536,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,5120,0.09970933198928833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,256,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,65536,10.070436265733507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,1536,0.20575200186835396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,10240,0.22638487815856934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,8192,0.019366222951147292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,5120,0.008354666332403818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,65536,0.25119198693169487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,8192,0.019673777951134574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,12288,0.0646488865216573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,256,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,256,0.0033902221669753394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,6144,0.09359111388524373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,5120,0.027451554934183758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,128,0.020081778367360432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,65536,0.0664915574921502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,3072,0.05488977829615275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,2048,0.00693866651919153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,768,0.1463528871536255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,512,0.006945778098371293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,256,0.007486222518814935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,2048,0.21378844314151338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,7168,0.010414222048388587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,10240,0.024473778075642053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,2560,0.03771200113826328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,1024,0.007786666353543599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,1536,0.004780444420046276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,768,0.009537777966923183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,51200,0.16724888483683267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,2560,0.005975111077229182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,51200,3.7261555989583335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,3584,0.03174577818976508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,51200,0.6683066685994467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,65536,0.044713778628243335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,12288,0.03150399857097202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,2048,0.006618666566080517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,12288,0.014950222439236112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,16384,0.8495155440436469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,128,0.003959111041492886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,5120,0.043624000416861645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,768,0.024897777371936377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,512,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,512,0.027958220905727808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,256,0.00675022229552269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,2048,0.024548444483015273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,65536,0.8643848631117078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,16384,0.056235558456844754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,512,0.008850666383902231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,65536,9.93786875406901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,16384,0.043190220991770424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,3072,0.022821333673265245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,6144,0.2943804528978136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,12288,0.05788977940877279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,1536,0.14323822657267252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,256,0.010661332971519895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,65536,0.20323467254638672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,768,0.01108444482088089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,5120,0.16613866223229304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,7168,0.020806221498383414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,8192,0.02094488839308421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,2048,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,128,0.016359110673268635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,32,0.01956622302532196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,32,0.15114311377207437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,7168,0.020167110694779288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,4096,0.04615999923812019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,6144,0.011542222566074796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,768,0.01828444500764211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,2048,0.0588888857099745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,10240,0.013640888863139682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,32,0.08775289191140069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,256,0.004435555388530095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,65536,9.080312093098959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,12288,0.023065777288542852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,51200,2.4743484920925565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,51200,0.11099821991390652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,65536,2.1949164072672525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,16384,0.06292711363898383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,12288,0.05136977632840475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,128,0.004089777668317159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,65536,0.07196178038914998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,64,0.005596444424655702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,64,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,768,0.02198933396074507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,16384,0.059798220793406166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,65536,0.08650488985909356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,12288,0.08554933468500774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,256,0.004831999954250124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,6144,0.09061244461271499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,3072,0.007833778030342525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,2048,0.007075555622577667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,256,0.012780444489585029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,3584,0.012645333177513547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,12288,0.3903840117984348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,128,0.004956444518433677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,5120,19.67257351345486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,128,0.007789333661397298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,8192,0.13624444272783068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,1536,0.10560266839133368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,2048,0.026086222794320848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,16384,4.869509379069011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,256,0.013274666335847644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,2048,0.0069466663731469055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,1024,0.021568889419237774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,128,0.025638222694396973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,512,0.01664533383316464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,32,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,6144,0.009584888815879822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,10240,0.4503653314378526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,65536,5.387241787380642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,64,0.00905600024594201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,3072,0.013243555194801755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,5120,0.008871999879678091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,32,0.02425066630045573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,1024,0.021384000778198242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,51200,0.06068266762627495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,51200,0.07037688626183404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,65536,0.0963360004954868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,4096,0.013601777454217276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,3072,0.006372444331645966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,256,0.025331555141343012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,65536,0.12246310710906982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,2560,0.011870221959220039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,65536,0.10957866907119751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,65536,0.5283599959479438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,256,0.003866666720973121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,1024,0.006544888847404056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,256,0.029543999168607924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,256,0.004150222159094281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,16384,0.9314728842841254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,7168,0.08417599731021458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,5120,0.5825982093811035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,5120,0.08571910858154297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,5120,0.016303999556435477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,768,0.028583112690183852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,32,0.0033777778347333274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,2560,0.13065600395202637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,5120,1.942263921101888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,256,0.024156444602542456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,128,0.00424799985355801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,512,0.004271111140648524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,1024,0.04954399996333652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,4096,0.14157156149546304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,1024,0.006057777752478917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,3072,0.05414666732152303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,768,0.003569777641031477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,16384,0.06396355231602986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,1536,0.00903733323017756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,16384,0.3111786577436659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,3072,0.06985421975453694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,10240,0.3270435598161486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,51200,0.061800890498691134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,768,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,128,0.005665777872006099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,16384,0.026455110973782007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,2560,0.007185777856243982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,2048,0.021407110823525324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,128,0.004506666627195147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,2560,0.11563911702897813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,12288,0.044891556104024254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,2560,0.040116445885764226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,16384,0.029988444513744775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,128,0.00507377791735861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,512,0.009306666751702627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,12288,0.29980354838901097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,65536,29.403584798177082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,2048,0.025193777349260118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,2048,0.0058871110280354815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,768,0.0038977778620190094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,768,0.007280000381999546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,65536,0.5296310848659939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,2048,0.007842666572994655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,65536,0.0680533316400316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,3584,0.013586666848924426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,2048,0.1851146618525187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,8192,0.19997333155737984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,128,0.004646222210592694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,1024,0.00924355536699295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,768,0.036887112591001726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,3072,0.010155555274751451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,512,0.0698231127527025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,768,0.026342221432262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,2048,0.03980711102485657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,4096,0.009314667019579146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,65536,0.191264894273546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,512,0.061795552571614586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,6144,1.344556490580241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,3072,0.09974933332867092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,3072,0.016840888394249808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,512,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,64,0.0029279999434947968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,7168,0.09458311398824056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,8192,0.031656887796190046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,16384,0.688718212975396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,32,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,32,0.24021156628926596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,3072,0.06810666455162896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,2560,0.28423645761277944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,64,0.008664000365469191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,5120,0.012835555606418185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,65536,0.13412533866034612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,1024,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,7168,0.022827555735905964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,5120,0.09099200036790635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,5120,0.09942933585908677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,3584,0.038627554972966514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,768,0.017850667238235474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,6144,0.39958755175272626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,65536,1.8128123813205297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,4096,0.010117332968446944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,5120,0.1351955599255032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,1024,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,8192,0.042897777424918286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,2048,0.018573333819707234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,2048,0.07391822338104248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,1024,0.01202666676706738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,65536,0.53656890657213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,7168,0.03004177742534214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,128,0.004234666625658671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,10240,0.03083911206987169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,32,0.031776888502968684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,256,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,3584,0.008164444731341468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,1536,0.004701333327425851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,6144,0.032891554964913264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,51200,0.3420728842417399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,3072,0.009711999860074785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,6144,0.01850222216712104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,64,0.0026177778426143858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,51200,1.8448399437798395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,3072,3.755118899875217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,32,0.010610666539933948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,51200,0.09132177962197198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,12288,0.021965333157115515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,768,0.009838222629494136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,12288,0.02758399976624383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,256,0.0036559998989105225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,2560,0.006249777972698212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,3584,0.0159262220064799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,10240,0.10503022538291083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,65536,1.5882497363620336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,65536,3.426100413004557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,3584,0.01851377718978458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,8192,0.019762666689025033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,16384,1.0989164776272244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,3584,0.15964889526367188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,10240,0.10721866952048408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,128,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,64,0.009178666604889764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,3584,0.017162667380438913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,16384,0.11387999852498372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,2048,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,10240,1.6287057664659288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,128,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,64,0.003167111012670729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,512,0.016971555021074083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,6144,0.025116443634033203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,4096,0.1549653344684177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,512,0.004940444396601783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,512,0.03433600068092346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,12288,2.100269317626953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,1536,0.0058346668051348785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,2048,0.05511733227305942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,4096,0.010808889236715106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,768,0.006996444529957241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,51200,3.504807154337565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,64,0.0107351111041175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,768,0.007176888485749562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,8192,0.008437333007653555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,16384,4.095447116427951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,2560,0.008004444340864817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,4096,0.008442666795518663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,51200,0.06942400005128649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,2048,0.007867555651399825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,128,0.014077333940400017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,16384,0.2509680059221056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,512,0.006770666688680649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,4096,3.312437269422743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,16384,0.16495022508833143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,1536,0.004956444518433677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,16384,0.15163289176093206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,512,2.374116473727756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,512,0.007850666840871176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,768,0.04180711176660326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,3072,0.019941333267423842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,7168,0.8376657697889539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,3584,0.16405155923631456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,3584,0.007850666840871176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,4096,0.06432888905207317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,128,0.027253333065244887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,3584,0.0229404436217414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,12288,0.15087466769748265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,10240,0.5718133184644911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,64,0.2026240030924479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,3584,0.008241777618726095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,128,0.0038782221575578055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,256,0.020184000333150227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,8192,0.01479822231663598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,256,0.005730666634109285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,128,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,12288,1.026942253112793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,7168,0.03824088970820109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,65536,0.34569422403971356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,1024,0.004546666724814309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,768,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,65536,0.07489155398474799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,6144,0.0070977773931291364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,1536,0.005146666533417172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,51200,0.5898364384969076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,512,0.004568888909286923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,6144,0.2669866614871555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,1024,0.0058462222417195635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,512,0.03428622086842855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,5120,13.47394053141276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,2560,0.007249777515729268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,64,0.010593777729405297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,768,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,3072,0.05987199809816149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,1024,0.045816000964906484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,256,0.053879112005233765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,5120,0.08124799860848321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,2048,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,65536,9.934034559461805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,12288,0.3923484484354655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,2048,0.007530666887760162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,512,0.005668444352017508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,6144,0.006234666539563074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,12288,0.03694577680693732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,8192,0.05486844314469231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,2560,0.08819466829299927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,512,0.005558222118351195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,8192,2.9454515245225696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,65536,0.16512266794840494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,4096,1.7969991895887587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,128,0.007637333538797166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,4096,0.011007111105653973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,3072,0.1086275577545166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,1024,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,1536,0.011624889241324531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,256,0.0032826666202810076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,5120,0.2315875424279107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,10240,0.050219555695851646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,2560,0.014309333430396186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,2048,0.15425599945916071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,4096,0.008729777402347988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,32,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,65536,1.618849754333496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,51200,1.3496551513671875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,64,0.0069475554757648045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,6144,0.024524443679385718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,8192,0.07325422101550631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,768,0.03173866536882188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,768,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,5120,3.060930675930447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,6144,0.11757066514756943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,16384,0.3352488941616482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,65536,0.4265013270907932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,65536,0.2445075511932373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,8192,0.038895110289255776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,5120,0.01684266659948561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,2560,0.04574133290184868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,12288,0.16271021631028917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,1536,0.006197333335876465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,3072,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,7168,0.08940800031026204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,768,0.008826666408114964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,3584,0.1806622213787503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,5120,0.1268791092766656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,4096,0.22266756163703072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,4096,0.04045244389110141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,1536,0.06404088603125678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,5120,0.012283555335468717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,16384,0.019632889164818656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,65536,0.21822578377193877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,32,0.04227111074659559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,4096,0.011901333100265928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,3072,0.030983110268910725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,1024,0.011036444041464063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,3072,0.10123288631439209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,7168,0.0269128895468182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,512,0.004923555586073133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,51200,1.288343111673991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,256,0.008904000123341879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,8192,0.016295111841625638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,256,0.008543999658690559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,16384,0.021299555897712708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,6144,0.13291911284128824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,4096,0.08005777994791667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,7168,0.014182221558358936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,512,0.006917333437336816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,3584,0.08902221918106079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,5120,0.028071112102932397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,3584,0.007320000065697565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,16384,0.3119893338945177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,8192,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,3072,0.00835111074977451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,1536,0.009098666409651438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,12288,0.04475377665625679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,4096,0.2397937774658203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,512,0.006962666908899943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,3584,2.438913769192166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,5120,0.015453333655993143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,7168,0.017833777599864535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,3072,0.007596444752481248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,512,0.0036355555057525635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,32,0.07881155278947619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,64,1.937704086303711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,1536,0.07508355379104614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,768,0.003564444267087512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,4096,0.03393066591686673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,65536,0.6483075353834364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,12288,0.015458666616015963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,256,0.03303111261791653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,7168,0.07369777891370985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,5120,0.03135111265712314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,5120,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,2560,0.005594666633341048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,1536,0.012271111210187277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,10240,0.19552711645762125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,6144,0.013952000273598565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,1024,0.04751555456055535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,512,0.01606044504377577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,7168,0.14994222588009304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,768,0.03597866826587253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,5120,0.3372773329416911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,51200,1.8802320692274306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,51200,0.07379200061162312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,32,0.0027964444210131965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,51200,1.6385609308878581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,10240,0.01533066729704539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,128,0.002999111182159848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,4096,0.3223715623219808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,1536,0.017248888810475666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,16384,0.08754844135708278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,2560,0.21817510657840303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,65536,0.6352648735046387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,128,0.004507555729813046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,1536,0.08702488740285237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,4096,0.03397511111365424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,10240,3.5235182444254556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,16384,2.2604196336534286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,512,0.03996977872318692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,1024,0.015189333094490899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,8192,0.15979377428690592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,65536,0.20703200499216715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,128,0.004399111287461387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,5120,0.06115733252631294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,16384,0.021807110971874658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,3072,0.009907555249002244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,10240,0.10014222065607707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,6144,0.27415288819207084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,4096,0.018924444913864136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,3584,0.012260444462299347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,3584,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,2560,0.08477333519193862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,7168,0.011759999725553723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,3072,0.013888888888888888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,3072,0.012824888858530255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,768,0.005678222411208683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,2560,0.009959999885823991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,128,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,4096,0.018481777773963082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,1536,0.010448888772063784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,2560,0.04986044433381823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,512,0.07366400294833712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,8192,0.06680978006786771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,12288,0.06401955419116549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,5120,0.08513155248430039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,256,0.005518222020732031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,1024,0.005195555587609609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,65536,1.8759119245741103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,1536,0.03665155503484938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,5120,0.04408444298638237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,2560,0.006744888921578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,1024,0.04295733239915636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,10240,0.9444586435953776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,16384,0.2890000078413221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,1024,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,1536,0.012432888978057437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,4096,0.08671555254194473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,3072,0.007528000407748752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,128,0.0028488888508743713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,32,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,4096,0.053914666175842285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,5120,0.015053333507643806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,256,0.005438222239414851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,8192,0.033301333586374916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,51200,0.03809600075085958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,768,0.003669333126809862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,1024,0.006552000012662675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,2048,0.008372444245550368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,10240,0.08087111181683011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,6144,1.4283946355183919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,8192,0.368505769305759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,768,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,3072,0.037865777810414634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,16384,0.7202293078104655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,16384,0.5649298032124838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,2560,0.016906667086813185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,3584,0.6491440137227377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,3072,0.12235377894507514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,5120,0.08159999714957343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,7168,0.015619556109110514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,768,0.009352000223265754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,16384,0.04344266653060913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,256,0.25895288255479604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,10240,0.6762417687310113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,512,0.012448888685968188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,12288,0.4274604320526123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,2560,0.007723555796676212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,2560,0.01718755563100179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,3584,0.008255999949243333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,256,0.018608000543382432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,5120,0.4286213450961643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,8192,0.016716443830066256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,2560,0.009551111194822524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,256,0.010287111004193624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,768,0.014746665954589844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,32,0.004978666702906291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,64,0.002560888934466574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,10240,0.015253333581818474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,512,0.004905777672926585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,1536,0.009002666506502364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,128,0.020232889387342665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,65536,0.5607235696580675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,51200,0.4229999913109674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,4096,0.07040088706546359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,1024,0.004212444441186057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,5120,0.23172354698181152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,32,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,4096,0.016206221448050607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,1024,0.02104533380932278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,2048,0.006686222222116258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,5120,0.10880355702506171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,10240,0.10790489117304485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,65536,0.14421155717637804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,2560,0.16538133886125353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,512,0.009278222090668148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,5120,0.27761867311265737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,2048,0.21447555224100748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,51200,0.055776887469821505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,3072,0.015617777903874716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,5120,0.05825333462821113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,3072,0.009397333694828881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,7168,0.0214008887608846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,4096,0.006057777752478917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,65536,24.885116577148438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,1536,0.01298577752378252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,51200,37.36761135525173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,10240,0.16954399479760063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,5120,0.0094035557574696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,128,0.005197333378924264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,8192,0.017952000101407368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,10240,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,256,0.007104000283612146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,65536,0.11363644070095485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,51200,1.67413510216607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,51200,0.019487111104859244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,32,0.00793866647614373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,3584,0.028766221470303003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,8192,0.020618667205174763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,512,0.003132444495956103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,2048,0.021312889125612047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,64,0.01368622233470281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,16384,0.014595554934607612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,2560,0.02414400047726101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,3072,0.01144977741771274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,5120,0.023286221755875483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,65536,9.872609456380209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,256,0.00507733349998792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,32,0.004931555440028508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,16384,3.195309321085612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,65536,0.6765999794006348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,2048,0.016880000631014507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,32,0.005245333330498801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,6144,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,12288,1.5285537507798936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,128,0.004287999951177173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,3072,0.0455644428730011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,3072,0.09346489111582439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,1024,0.0052382221652401825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,2048,0.02104622291194068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,3072,0.056405332353379994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,4096,0.008014221986134848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,2560,0.007631999750932057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,512,0.03550666570663452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,768,0.02177511155605316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,16384,0.00944622192117903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,1024,0.015011555618709989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,3072,0.011986666255527072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,4096,0.41106221410963273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,3584,0.3182355562845866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,5120,0.060862223307291664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,5120,0.16525867250230578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,3584,0.012416889270146688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,7168,0.03033777740266588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,10240,0.024005333582560223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,3584,0.010130667024188572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,2048,0.014354666074117025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,2560,0.059049778514438205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,10240,0.10134577751159668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,2560,0.041936000188191734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,512,0.005665777872006099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,2048,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,12288,0.021644444929228887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,3584,0.00869422240389718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,512,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,4096,0.022131555610232886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,16384,0.09850666920344035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,4096,0.07143377595477633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,128,0.05856800079345703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,8192,0.2689395480685764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,10240,0.018592889110247295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,65536,0.10702755716111924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,8192,0.02148355543613434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,512,0.0061048890153567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,8192,0.12034132745530869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,10240,0.024676443801985845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,128,0.011391999820868174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,7168,1.2887004216512044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,1536,0.006454222318198945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,32,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,768,0.027990221977233887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,12288,0.4685813585917155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,7168,0.10099466641743977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,32,0.015170666906568738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,8192,0.028246220615175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,4096,0.17674577236175537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,3072,0.0151582227812873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,7168,0.10938933160569932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,512,0.08526221911112468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,2048,0.17795111073387992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,10240,0.12636266814337835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,51200,0.07853244410620795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,32,0.012088889049159156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,32,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,1024,0.04508089025815328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,4096,0.04375999834802416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,5120,0.03215822246339586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,6144,0.026661333110597398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,6144,0.1065066655476888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,2560,0.007037333316273159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,768,0.004030222280157937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,51200,0.012889777620633444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,7168,0.06619022289911906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,6144,0.010612444745169746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,128,0.0041333333485656315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,10240,0.13465689288245306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,256,0.006865777903132969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,3584,0.01239111108912362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,8192,0.08318399720721774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,128,0.013355555633703867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,51200,0.008445333275530074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,3072,0.12161600589752197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,7168,0.02328444520632426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,2048,0.016196444630622864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,65536,0.11638311545054118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,6144,0.020661332541041903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,2048,0.015793777174419828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,65536,0.11925244331359863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,3072,0.014570666684044732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,3584,0.025036444266637165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,3584,0.012097777591811286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,768,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,1536,0.06813422176573011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,12288,0.05972177452511258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,6144,0.031975110371907554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,6144,0.01090400003724628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,2048,0.013343999783198038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,3072,0.009617778162161509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,51200,0.12528710895114475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,51200,0.6535786522759331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,768,0.0042275554604000514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,12288,0.024865777956114873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,768,0.007461333440409766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,512,0.007792000141408708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,7168,0.3370070987277561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,10240,0.19098488489786783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,10240,0.04862044586075676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,1536,0.044165333112080894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,8192,0.41016266081068253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,64,0.012811555630630918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,64,0.008762666748629676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,5120,0.012240000069141388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,3584,0.15802578131357828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,768,0.0413680010371738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,1024,0.14301421907213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,5120,0.030460443761613634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,256,0.003313777761326896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,64,0.005168888717889786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,3072,0.03404266635576884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,1024,0.010026666853162978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,1024,0.004879111217127906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,8192,0.02271733350223965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,768,0.004078222231732475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,65536,0.06849066416422527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,768,0.008002666963471307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,2560,0.007673777639865875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,4096,0.006752000086837345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,3072,0.006888000087605582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,1024,0.0058613332609335584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,16384,0.37163911925421816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,7168,0.013212444053755866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,12288,0.018339556124475267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,51200,0.23408354653252494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,16384,0.23682755894131133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,32,0.0026604444202449587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,128,0.03020711077584161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,51200,7.5774019029405375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,1536,0.006174222048785951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,512,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,65536,0.08873777919345432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,1536,0.5005742179022895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,6144,0.012517333030700684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,7168,0.023588443795839947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,16384,0.023060444328520033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,12288,0.021116443806224402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,2560,0.007713778151406183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,512,0.003539555602603488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,8192,1.2261955473158095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,256,0.00378577783703804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,5120,0.011373333632946014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,2048,0.014440889159838358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,8192,0.023752000596788194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,2560,0.005964444329341252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,2048,0.022258667482270136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,16384,0.29345777299669057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,8192,0.1340577734841241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,768,0.014706666270891825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,16384,0.49210845099555117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,2048,0.024305777417288885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,7168,0.2600773440466987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,2048,0.05552355448404948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,3072,0.016504888733228047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,32,0.0030533334033356774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,2560,0.06082489093144735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,256,0.0075146663520071245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,64,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,512,0.008092444803979661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,10240,0.03218400147226121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,512,0.02012266715367635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,2048,0.006695999867386288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,1024,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,8192,0.7646222114562988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,6144,1.7438071568806965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,128,0.018732445107565988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,128,0.01651733285850949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,65536,0.03334666623009576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,768,0.006120000034570694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,16384,0.07830844322840373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,4096,0.022069333328141108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,4096,0.0693991118007236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,512,0.006078222145636876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,8192,0.08346666892369588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,32,0.03639199998643663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,2048,0.006940444310506185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,65536,0.2986231115129259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,5120,0.11201955212487115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,1536,0.04932622114817301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,3584,0.47696442074245876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,768,0.048927999205059476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,8192,0.019931554794311523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,5120,0.15437600347730848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,51200,0.11003822088241577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,2560,0.017921777235137094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,128,0.024689777029885188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,65536,1.4490906397501628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,16384,0.23705244064331055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,10240,0.6077209048800999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,16384,1.3732648425632055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,16384,0.17094133959876168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,256,0.00398488880859481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,2048,0.009279111193286048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,3584,0.024588444166713293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,65536,0.013267555170589022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,7168,0.04222933451334635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,5120,0.17880000008477104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,6144,0.02199466692076789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,6144,0.06425955560472277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,1024,0.06400533517201741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,1536,0.011642666326628791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,8192,0.09380622042549981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,8192,0.0644764436615838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,128,0.007485333416197035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,51200,23.09811570909288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,2048,0.009325332939624786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,512,0.005626666463083691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,7168,0.017308443784713745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,3584,0.045941332976023354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,65536,0.3348373307122125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,4096,0.038763556215498186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,5120,0.008871999879678091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,16384,0.03151822090148926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,3584,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,7168,0.032877332634396024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,8192,0.04310933417744107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,12288,0.3073066605461968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,7168,0.4631004333496094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,16384,0.022224000758594934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,1024,0.008904889225959778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,768,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,3072,0.12642044491238064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,65536,24.919547186957463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,7168,0.01258133351802826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,5120,0.10143555535210504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,8192,0.02432800001568264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,4096,0.007071110937330458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,512,0.005975111077229182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,1536,0.010025777750545079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,51200,0.9687680138481988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,128,0.0037626665499475268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,12288,0.047297779056761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,16384,5.825941297743056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,768,0.01354311075475481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,1024,0.013231111069520315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,16384,0.023456000619464453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,3584,0.008932444784376357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,3584,0.09458577632904053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,51200,1.2872737248738606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,65536,0.1283751063876682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,512,0.004680888934267892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,51200,18.156667073567707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,128,0.004514666481150521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,12288,0.43328176604376895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,5120,2.1934471130371094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,7168,0.018381334013409086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,4096,1.469127125210232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,64,0.0041893331540955436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,51200,0.07517244418462117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,256,0.0059573331640826324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,1024,0.013031110995345645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,2048,0.05179288983345032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,256,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,12288,0.17066489325629342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,64,0.025455999705526564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,5120,0.026515555050637987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,768,0.005679111099905438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,8192,0.01458044515715705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,768,0.00600888869828648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,16384,0.06776533524195354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,8192,0.021253334151373968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,2560,0.024004444479942322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,3584,0.018226666582955253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,4096,0.03017866611480713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,3584,0.20436267058054605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,3072,0.007479999628331926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,768,0.03170399864514669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,65536,0.7047004699707031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,3072,0.021928000781271193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,1536,0.006519111080302133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,2560,0.007419555551475949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,8192,2.956487019856771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,10240,0.08117333385679457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,256,0.0032791110376516977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,65536,12.967855665418837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,4096,0.012668444050682915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,65536,1.7160266240437825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,10240,0.03534488876660665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,3584,0.034902221626705594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,5120,0.013829333086808523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,8192,0.546928882598877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,256,0.022424888279702928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,65536,0.0971475574705336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,10240,0.49434132046169704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,64,0.017716444200939603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,2560,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,51200,0.9723040262858073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,7168,0.013634666800498962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,8192,0.1743555598788791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,10240,3.8699518839518228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,7168,0.21990933683183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,768,0.006980444822046492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,256,0.02307022280163235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,8192,0.017663111289342243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,8192,4.905799865722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,12288,0.024522667129834492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,12288,0.017709333035680983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,1024,0.0038248888320393036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,65536,0.07631111145019531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,128,0.0029102222373088202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,1536,0.011699555648697747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,64,0.08262844218148126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,16384,0.03791733251677619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,65536,0.1793217791451348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,65536,0.4616070853339301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,51200,0.2931804392072889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,32,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,32,0.0053119998839166425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,51200,0.16807377338409424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,1024,0.010271111296282874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,4096,0.03404266635576884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,6144,0.06475733386145698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,8192,0.15011555618709987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,2048,0.009850666754775578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,5120,0.03500533435079787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,65536,0.08349600103166367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,51200,0.18702399730682373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,1024,0.06184533569547865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,2048,0.018209778600268893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,64,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,1536,0.010275555153687796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,12288,0.2032639980316162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,64,0.010262221925788457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,2560,0.01499111122555203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,1536,0.012239110966523489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,512,0.00815466625822915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,2048,0.042266666889190674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,5120,0.01479022204875946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,16384,0.6999555693732368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,512,0.0703528920809428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,1536,0.024455999334653217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,10240,0.024957334001859028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,65536,2.1305857764350042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,2560,0.02086044516828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,3584,0.022016889519161646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,64,0.012764444781674279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,5120,0.007287110719415877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,51200,0.21804800298478869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,2048,0.006762666834725275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,1024,0.02295911146534814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,12288,0.19142844941880968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,8192,0.12704355186886257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,12288,0.09569511148664687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,128,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,3072,0.1381982167561849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,2048,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,3584,0.023969777756267126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,10240,0.19236355357699922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,256,0.005152889009979037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,2048,0.007296000089910295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,5120,0.006736889067623351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,4096,0.3057742118835449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,512,0.020655999581019085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,768,0.010514666636784872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,512,0.0033244445092148254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,2048,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,12288,0.030144890149434406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,2048,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,5120,0.02901866700914171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,256,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,512,0.007988444632954067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,2560,0.1121786700354682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,51200,2.818637424045139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,2560,0.18253866831461588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,5120,0.031368888086742826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,7168,5.046663920084636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,5120,0.019867555962668527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,51200,0.08471644586986965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,2048,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,65536,2.7363084157307944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,65536,0.018283555905024212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,2560,0.13515733347998724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,64,0.00654666663871871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,8192,0.030464887619018555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,12288,0.048856000105539955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,512,0.008972444468074376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,65536,0.3272808922661675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,51200,1.2257937325371635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,4096,0.7877404424879285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,7168,0.07254222366544935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,2560,0.008043555749787224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,6144,0.0181057784292433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,7168,0.34448888566758895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,16384,0.48592355516221786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,65536,0.3346889019012451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,7168,0.0064186664919058485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,8192,0.088392886850569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,2560,0.006501333167155583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,256,0.0037582222786214617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,1024,0.012804444465372296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,65536,0.28311199612087673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,4096,0.016167999969588388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,3584,0.01942222151491377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,10240,0.007304888632562425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,12288,4.921587626139323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,256,0.004596444467703502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,10240,0.019337778290112812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,12288,0.023975999818907842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,51200,0.1364533371395535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,4096,0.007439110842016008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,2048,0.07723466555277507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,3584,0.011095110740926532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,65536,1.3311786651611328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,3072,0.02147377696302202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,65536,1.881715562608507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,32,0.004994666824738185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,768,0.00699733363257514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,3584,0.014506666196717156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,1024,0.012685333689053854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,5120,0.013521778086821238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,64,0.003654222107595868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,12288,0.41257601314120823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,10240,0.027652444110976324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,128,0.0026968889352348116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,1536,0.01201600001917945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,2048,0.006983111302057902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,4096,0.008399110701349046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,2048,0.006090666684839461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,8192,0.0950746668709649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,5120,0.1544746690326267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,1024,0.003926222109132343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,1536,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,1536,0.007278222176763747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,10240,0.017256889078352187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,512,0.06748977634641859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,5120,0.017252445220947266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,12288,0.24217687712775338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,3072,0.017049777838918898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,1536,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,7168,0.07571378019120958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,8192,0.01646399994691213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,6144,0.0232968893316057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,6144,0.7512728903028699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,6144,0.030622220701641504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,3072,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,65536,0.6851698027716743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,12288,0.6127893129984537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,2048,0.04740266667471992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,10240,1.0916897455851238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,16384,0.06522755490409003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,1024,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,1024,0.015263110399246216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,32,0.0074133334888352295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,256,0.005159111072619756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,32,0.002528888897763358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,51200,0.050498667690489024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,768,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,5120,0.009438222481144799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,768,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,64,0.006008000009589725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,16384,0.022686221533351477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,8192,0.04927466644181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,10240,0.022594667143291895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,2048,0.09463289048936631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,1024,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,10240,0.20360888375176323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,256,0.003097777772280905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,6144,0.03130577670203315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,512,0.02591111097070906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,2048,0.010628444453080496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,128,0.03920355439186096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,4096,0.008616888688670265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,8192,0.022042666872342426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,3072,0.039735111925337054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,5120,0.01128799964984258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,10240,0.10161777999666001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,5120,0.006315555423498154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,10240,0.02572088936964671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,6144,0.018170666363504197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,1536,0.04784622126155429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,7168,0.017529777354664274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,6144,1.1582959493001301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,64,0.0038240001433425476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,6144,0.010256888965765635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,768,0.3841431140899658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,4096,0.04706044329537285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,6144,2.5278231302897134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,32,0.0032524443748924467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,256,0.013819555441538492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,7168,0.02590222160021464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,64,0.011147555377748279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,32,0.0034293333689371743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,1024,0.033632887734307185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,12288,0.7665831247965494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,512,0.005032888717121548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,12288,0.013424889081054263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,1536,0.4357573456234402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,2048,0.04612088865704007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,64,0.002653333254986339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,4096,0.02338933282428318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,8192,0.06614755259619819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,12288,0.35290132628546816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,2048,0.011839110818174152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,128,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,768,0.050422221422195435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,768,0.03980266716745164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,3584,0.01108977778090371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,65536,2.8055163489447703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,5120,0.010840888652536603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,6144,0.0402817792362637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,768,0.005897777775923411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,4096,0.10911555422676934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,51200,1.2465733422173393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,10240,0.5524160067240397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,16384,0.024880889389250014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,3584,0.10204355584250556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,65536,0.9235146840413412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,3584,0.29796889093187123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,32,0.00423555572827657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,4096,0.4686711099412706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,2560,0.07561510801315308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,64,0.017302221722073026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,3584,0.03348533312479655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,51200,0.07400977611541748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,6144,0.03370489014519586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,64,0.019680000013775296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,64,0.002797333316670524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,256,0.0029208889851967492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,51200,0.11418755849202473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,65536,0.06821066803402372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,2048,0.08994844224717881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,2048,0.009259555902745988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,7168,0.21657866901821562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,64,0.02123911182085673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,7168,0.24933332867092559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,32,0.010053333308961656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,12288,0.010209778116808997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,10240,0.12388443946838379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,1024,0.026533333791626826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,3584,0.02230577833122677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,2048,0.007515555454625024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,65536,0.748494201236301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,128,0.008236444658703273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,768,0.04366755485534668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,16384,4.8631642659505205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,768,0.018162666095627677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,2048,0.016352888610627916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,12288,0.01051999959680769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,7168,0.01274400038851632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,12288,0.14542667071024576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,3072,0.036622222926881574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,3584,0.011205332974592844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,4096,0.00830844458606508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,8192,0.12424622641669379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,65536,5.1463877360026045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,5120,0.05217777689297994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,512,0.11270399888356526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,1024,0.00573244442542394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,3584,0.011011555790901184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,1536,0.00563822231358952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,10240,0.031974222924974226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,3072,0.02030666669209798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,1024,0.019609777463806998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,2560,0.011791999969217511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,16384,0.03293422195646498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,10240,0.021910222040282354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,65536,0.5893217722574869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,3072,0.00793688909875022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,7168,0.04562577936384413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,12288,0.036115556955337524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,64,0.014193778236707052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,65536,0.16529511080847845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,128,0.016248888439602323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,51200,0.35884621408250594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,1536,0.009502222140630087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,3072,0.018869333797030978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,64,0.017804443836212158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,12288,0.042484445704354175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,512,0.05617066886689928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,3584,0.04316088888380262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,1024,0.013103110922707452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,1536,0.041638222005632185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,3072,0.2751777701907688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,1024,0.007997333175606197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,3584,0.0070497774415545994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,16384,0.022672888305452134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,12288,0.09469155470530193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,10240,1.3805440266927083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,5120,0.08731200297673543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,8192,0.06612533330917358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,8192,0.1297875510321723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,51200,0.10391288995742798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,256,0.004805333498451444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,256,0.007178666690985362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,16384,0.05319911241531372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,5120,0.138629330529107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,4096,0.008488000267081791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,768,0.07824622260199653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,16384,0.1306017769707574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,768,0.008202667037645975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,1024,0.048009776406817965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,512,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,768,0.01587466730011834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,256,0.003137777869900068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,6144,0.036320888333850436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,16384,0.28990311092800564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,1536,0.006054222169849608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,64,0.00721955547730128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,4096,0.05860977702670627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,2560,0.01900533338387807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,12288,0.040419555372662015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,512,0.014336888988812765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,3584,0.10794666740629409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,6144,0.3072311083475749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,512,0.005676444619894028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,12288,0.36649245686001247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,1024,0.004927111168702443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,16384,0.16592355569203696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,10240,0.014866666661368476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,64,0.004286222159862518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,3584,0.4715253512064616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,768,0.004108444270160464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,51200,42.40155029296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,4096,0.007272889216740926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,12288,0.04079111085997687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,768,0.5391786893208822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,32,0.24662666850619844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,8192,0.02204888893498315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,64,0.005715555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,10240,0.010278222461541494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,2560,0.19865155220031738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,3072,0.009000000026490953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,65536,0.06820088624954224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,512,0.0058124446206622655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,768,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,3072,0.062295112344953746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,10240,0.06161510944366455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,64,0.006364444477690591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,2560,0.006881777611043718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,65536,0.022427555587556627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,1536,0.012441777520709567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,768,0.016914667354689706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,1024,0.017494221528371174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,65536,0.9690613216824002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,5120,0.07932800054550171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,5120,1.461310174730089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,2048,0.07080622514088948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,4096,0.010830221904648675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,1536,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,6144,0.04119644562403361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,2048,0.010431110858917236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,32,0.00309599998096625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,6144,0.0148053334818946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,16384,0.04139022363556756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,3584,0.009903111391597325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,1536,0.00796622203456031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,768,0.03786755601565043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,16384,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,64,0.0029200000895394217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,6144,0.010925333533022137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,12288,0.2698613272772895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,65536,0.04001955522431268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,3072,0.007453333172533248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,10240,0.013502221968438892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,8192,0.02814399864938524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,5120,0.012056888805495368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,5120,0.04071111149258084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,65536,0.5250319904751247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,2048,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,7168,0.00868444475862715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,64,0.009399111072222391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,64,0.02347911066479153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,1536,0.0070675553547011475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,16384,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,51200,0.07459644476572673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,6144,1.5002418094211156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,256,0.006810666786299811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,65536,0.9979786343044705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,1536,0.005668444352017508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,512,0.03205510973930359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,51200,0.05402400096257528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,768,0.007315555380450354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,2048,0.005921777751710679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,3072,0.030504888958401147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,16384,0.7683004273308648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,256,0.0032791110376516977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,3584,0.02360711163944668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,4096,0.021753778060277302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,32,0.0058755555914507965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,16384,1.2460418277316623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,6144,0.03271377748913235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,51200,0.6009475390116373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,3072,0.02442488902144962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,3072,0.034852445125579834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,2560,0.006223111102978389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,512,0.010118222071064843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,8192,0.1605502234564887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,5120,0.009506666825877296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,65536,0.14999200238121882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,1024,0.010117332968446944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,12288,0.01538222200340695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,16384,0.02104266650146908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,128,0.006340444501903322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,768,0.006663110935025745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,2560,0.08436622222264607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,16384,0.03872266742918227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,1536,0.0094355551732911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,6144,0.16318489445580378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,3072,0.06650933292176989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,3072,0.06685511271158855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,16384,0.03590933481852213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,12288,0.030677331818474665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,65536,0.06062666575113932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,51200,0.4377342330084906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,3072,0.023649776975313824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,768,0.005993777679072485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,32,0.003244444520937072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,12288,0.10928977860344781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,2048,0.6823609140184191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,256,0.009853333234786987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,5120,0.006274666637182236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,8192,0.06025422281689114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,256,0.06421066655053033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,128,0.011022222538789114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,7168,0.06301155355241564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,7168,1.4487804836697047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,1536,0.05606222152709961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,12288,10.18606482611762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,12288,0.2523182233174642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,51200,0.2598115603129069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,3584,0.017119111286269296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,2560,0.06447822517818876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,32,0.007371555599901412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,65536,3.1190488603379993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,64,0.0033235556135574975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,8192,0.026463111241658527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,128,0.0034577778230110803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,256,0.008952889177534316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,64,0.003726222034957674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,3584,0.057562669118245445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,5120,0.19983377721574572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,256,0.009764444496896531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,512,0.005671111245950063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,32,0.016210666961140104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,6144,0.18199733893076578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,12288,0.3542017671797011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,51200,0.06156000163820055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,768,0.006498666687144174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,12288,0.06734044684304132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,1536,0.4792328940497504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,768,0.026759111218982275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,64,0.006771555377377405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,10240,0.02385777731736501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,7168,0.057462222046322294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,64,0.004960888789759742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,1024,0.02050488856103685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,8192,0.023553777072164748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,256,0.004519111166397731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,5120,0.13382932874891493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,4096,0.10632266600926717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,2048,0.014184888866212634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,32,0.03426488902833726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,3072,0.11683289210001628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,6144,0.03093155556254917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,65536,3.526408937242296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,16384,4.5228801303439665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,1536,0.012235555383894177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,64,0.006508444332414203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,16384,1.6859857771131728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,128,0.007394666473070781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,64,0.00693866651919153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,256,0.04637066523234049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,256,0.02658933401107788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,5120,0.06833422183990479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,51200,0.16105600198109946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,2048,0.06193866994645861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,6144,0.33823198742336696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,65536,0.9206844965616862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,7168,0.008634666601816813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,8192,0.035190221336152815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,5120,0.02126311096880171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,512,0.013302221894264221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,256,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,6144,0.03661777906947666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,2560,0.006993778049945831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,1536,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,128,0.005138666679461797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,2048,0.02964444292916192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,8192,0.0214382227924135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,4096,0.07486666573418511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,4096,0.03285333183076646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,1024,0.08269688818189833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,10240,0.1795191102557712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,128,0.00851644410027398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,5120,2.441350301106771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,128,0.0107342220014996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,128,0.00267999991774559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,12288,0.055083556307686694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,5120,0.06672355863783094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,51200,0.042499555481804736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,7168,0.2997208966149224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,1536,0.08919111225340103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,7168,0.03153866529464722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,16384,0.2963893413543701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,128,0.024774221910370722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,4096,0.04400888747639126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,16384,0.127329773373074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,128,0.005145777844720417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,6144,0.016591111818949383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,3072,0.02646044393380483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,768,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,128,0.026399999856948853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,64,0.006572444405820634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,128,0.007788444558779399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,3072,0.010294222169452244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,1024,0.020872889293564692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,8192,0.2817777792612712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,51200,0.06020444631576538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,10240,0.016376000311639573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,3584,0.01591822173860338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,768,0.19086221853892008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,2560,0.26658667458428276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,2048,0.006909333169460297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,5120,0.009674666656388177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,5120,0.02344977855682373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,16384,0.04651022288534376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,1536,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,6144,0.025802666942278545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,51200,0.11638311545054118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,3072,0.038386666112475924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,65536,0.066012442111969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,2048,0.009583111438486312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,2048,0.005333333379692501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,51200,0.5341591305202907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,32,0.007174222005738153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,512,0.022072000635994807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,12288,0.22413688235812715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,8192,0.026007999976476032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,10240,0.021041777398851182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,256,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,5120,0.025664889150195654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,10240,0.08558933602439033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,256,0.005761777775155173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,32,0.0026337777574857077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,3072,0.02937955657641093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,1536,0.006167999986145232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,1024,0.007855110698276097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,51200,0.11946222517225479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,65536,0.23911465538872612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,768,0.004618666652176115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,12288,0.19363821877373588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,512,0.05187111099561056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,8192,0.5138044357299805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,3584,0.007864889171388414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,2560,0.04074755642149184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,256,0.01310488912794325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,10240,0.08711910910076565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,256,0.004687110996908612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,3584,0.014789332946141561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,7168,0.03421511252721151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,1536,0.010601777997281818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,512,0.08481244246164958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,1024,0.038365334272384644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,12288,0.36283111572265625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,1536,0.004990222139490975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,128,0.1480497784084744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,8192,0.0357528891828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,3072,0.021586666504542034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,64,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,1024,0.007183111376232571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,51200,0.05697866943147448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,65536,6.974159240722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,10240,0.2804311116536458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,512,0.02754844393995073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,128,0.011240888800885944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,16384,0.8191226853264703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,16384,0.1372391117943658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,1536,0.013904000322024027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,512,0.027306665976842243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,3584,0.010330666270520952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,6144,0.5581706894768609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,5120,0.041441778341929116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,768,0.005321777943107817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,16384,0.0977511141035292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,51200,0.04952444301711189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,5120,0.011901333100265928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,32,0.002993777808215883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,128,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,3584,0.39624977111816406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,1536,0.009587555295891231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,32,0.0030951110853089225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,1024,0.01183377785815133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,32,0.0032257777121331957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,7168,0.01311466677321328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,2048,0.04714666803677877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,4096,0.12416177325778538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,2048,0.07899110847049289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,65536,0.4284613397386339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,4096,0.024951110283533733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,3584,0.06156622038947212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,1024,0.007521777517265744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,3584,0.006268444574541516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,512,0.014552889598740472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,8192,0.2422888808780246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,6144,0.053545776340696544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,51200,0.8774933285183377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,64,0.02497600018978119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,32,0.01094133324093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,6144,0.017121778594122995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,65536,26.859217325846355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,10240,0.41109776496887207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,64,0.018623110320832994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,512,0.03798933492766486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,5120,0.013062222136391534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,768,0.01696711116366916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,128,0.09163110786014134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,256,0.0038071109188927543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,2048,0.034304890367719866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,12288,0.05372533202171326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,32,0.008532444636027018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,256,0.002899555489420891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,128,0.002700444517864121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,256,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,7168,1.3310897615220811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,12288,0.0396728879875607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,12288,0.019271110494931538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,2560,0.023787554767396715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,256,0.08099110921223958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,2560,0.015029332704014249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,8192,0.04838844471507602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,10240,0.08760533067915176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,5120,0.07429511017269559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,7168,0.014094221923086377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,256,0.0029537777105967202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,256,0.016571554872724745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,2048,0.011067555182509951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,10240,0.3268800046708849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,8192,0.036441776487562395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,3072,0.006320000108745363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,512,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,3584,0.28025688065422905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,2560,0.008054222497675154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,2048,0.005012444323963589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,5120,0.007438222567240397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,2560,0.006650666809744305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,256,0.006375999914275275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,16384,11.566283331976997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,3072,0.2573022312588162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,12288,0.008906666603353288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,5120,0.013827555709415011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,5120,0.021178666088316176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,256,0.015634665886561077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,1024,0.0127688886390792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,1536,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,3584,0.016383111476898193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,12288,0.07654666900634766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,2048,0.016540444559521146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,10240,0.16941955354478624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,8192,0.5037244690789117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,1536,0.034661332766215004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,2048,0.11363200346628825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,1536,0.04986133178075155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,2048,0.010879110958841113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,65536,0.07259022527270846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,7168,0.045351998673544995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,1024,0.2978302107916938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,12288,0.5871520042419434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,65536,1.515718248155382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,5120,0.013464888764752282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,3072,0.049505776829189725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,6144,0.013083555632167392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,12288,0.027144889036814373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,7168,0.027482666903071936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,8192,0.02776622275511424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,8192,0.12655911180708143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,3072,0.0069004446268081665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,512,0.026927999324268762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,3584,0.040214220682779946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,3584,0.01664622293578254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,4096,0.05211288730303446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,3584,0.026076444321208533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,10240,0.2476737764146593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,4096,0.07595555649863349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,51200,0.06991377804014418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,10240,0.16039466857910156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,51200,0.06111288732952542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,10240,0.013762666947311826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,1536,0.007419555551475949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,51200,17.656021118164062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,2560,0.08984977669186062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,256,0.004435555388530095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,7168,0.04773333337571886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,3072,0.006264888991912206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,8192,2.7667164272732205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,2048,0.04427822099791633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,128,0.005431111074156231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,4096,0.014846222268210517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,2560,0.04972266819741991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,3072,0.0063031112982167145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,3072,0.01275288893116845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,3584,0.02522666586769952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,5120,0.01408088869518704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,32,0.007681777907742395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,768,0.0037768888804647657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,3584,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,128,0.01407199932469262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,1024,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,32,0.0074755557709270054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,12288,0.08033244477377997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,16384,0.1173822217517429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,2048,0.01924444403913286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,51200,0.12300977441999648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,65536,0.18571110566457114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,1536,0.005573333137565189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,32,0.008235555556085374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,128,0.0028115556471877625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,10240,0.9260969161987305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,768,0.006179555422729916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,12288,0.10508177677790324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,2560,0.007771555748250749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,768,0.0064897777305708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,51200,2.518674638536241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,3072,0.023524444964196947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,1024,0.004852444347408083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,128,0.0034266666819651923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,51200,0.13744621806674534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,3072,0.013569778038395775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,2048,0.010877333581447601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,51200,0.0464933349026574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,3072,0.007626666790909237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,7168,0.047876444127824574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,128,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,3584,0.23753423160976836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,256,0.0047866664826869965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,3584,0.013796444568369122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,32,0.01126933346192042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,256,0.004098666624890434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,2048,0.011035555766688453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,2560,0.025983999172846477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,512,0.003552888830502828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,4096,0.00960533320903778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,256,0.02795377704832289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,768,0.016456888781653512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,12288,2.331264919704861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,3072,0.00980444418059455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,65536,0.03030844529469808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,16384,0.12266577614678277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,8192,0.4304506778717041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,10240,0.027082666754722595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,3072,0.2548248767852783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,6144,0.4517599741617839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,2560,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,10240,0.007102222078376346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,16384,0.02723466687732273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,32,0.027982221709357366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,16384,1.2138995064629448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,2048,0.025514667232831318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,256,0.004878222114510006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,51200,1.2289057837592232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,64,0.010011555420027839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,8192,0.2910924487643772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,3072,0.0343413319852617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,4096,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,512,0.003351111171974076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,2048,0.03354133168856303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,3072,0.11060177617602879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,1536,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,16384,0.42304354243808323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,256,0.0037706668178240457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,2048,0.008620444271299574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,1536,0.016716443830066256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,4096,0.018045332696702745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,3584,0.1347786717944675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,16384,0.03249866763750712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,2048,0.2067191070980496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,7168,0.25952355066935223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,3584,0.011734222372372946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,3072,0.008798222574922774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,4096,0.014264000786675347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,51200,0.7631742159525553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,2048,0.05609333515167236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,256,0.055564443270365395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,16384,0.06624355581071642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,16384,0.03418488966094123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,65536,0.41611999935574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,512,0.015927111109097798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,2560,0.012823999755912356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,1024,0.022256000174416438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,2048,0.02560088866286808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,8192,0.02009600069787767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,3584,0.029089778661727905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,4096,0.006213333457708359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,51200,0.08011288775338067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,8192,0.011552000211344825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,64,0.0031226666437255014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,6144,1.6477129194471571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,8192,0.034889777501424156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,2560,0.045590221881866455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,3072,0.02529955572552151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,1024,0.032278223170174494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,51200,0.02477866742346022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,768,0.009191110730171204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,3072,0.01997066703107622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,16384,0.29751022656758624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,65536,0.42203201187981504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,5120,0.04203644394874573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,6144,0.021762667430771723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,768,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,2560,0.007852444218264686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,3584,0.2835297849443224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,10240,0.03529511226548089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,1024,0.021822222405009802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,12288,1.4407786263359919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,2560,0.09367911020914714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,4096,0.13421777884165445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,512,0.004825777891609403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,3584,0.12377599875132243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,768,0.004490666505363253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,7168,1.1983022689819336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,2048,0.01667022208372752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,1024,0.008624888956546783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,64,0.05338489015897115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,7168,0.25139111942715114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,16384,2.472621281941732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,7168,3.285869386461046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,2048,0.6785546408759223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,16384,0.025920889443821375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,32,0.0033119999700122406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,2048,0.017400888933075797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,51200,0.19548622767130533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,10240,0.042372445265452065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,12288,0.2690311008029514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,64,0.0027253333893087176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,4096,0.01870044403605991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,5120,0.05453333258628845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,51200,0.10219555430942112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,1536,0.00814577771557702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,768,0.005115555392371284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,2048,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,128,0.003050666716363695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,768,0.008396444221337637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,768,0.04056444433000352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,512,0.00490399988161193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,2560,0.006133333262470033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,65536,1.2908044391208224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,16384,4.107555389404297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,10240,0.07768355475531684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,6144,0.14208443959554037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,16384,0.01943644384543101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,12288,0.07581155829959445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,6144,1.578543980916341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,7168,0.015456888410780164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,10240,0.013950222068362765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,65536,0.08457777897516887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,10240,0.06475910875532362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,1536,0.005498666730191972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,256,0.012631110846996307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,1536,0.013629333012633853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,65536,1.1617404090033638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,16384,0.06915555397669475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,768,0.11107911003960504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,7168,0.21548799673716226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,1536,0.010355555348926121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,3072,0.010311110979980892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,7168,0.016889777448442247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,64,0.060806221432156034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,8192,0.010433778166770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,128,0.003428444473279847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,1536,0.012310222619109683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,256,0.006493333313200209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,7168,0.1956826713350084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,10240,7.193636576334636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,3584,0.009666666388511658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,512,0.0037377778854635027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,64,0.003579555700222651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,1536,0.027238221632109746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,1024,0.003891555385457145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,8192,0.6222266621059841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,12288,0.010593777729405297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,64,0.0026213334252436957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,12288,0.05954222546683418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,16384,0.020634666085243225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,5120,0.2041217750973172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,32,0.0033617777129014335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,12288,0.08731022146013047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,2048,0.04043822156058417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,6144,0.012739555703269111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,65536,0.0674462252193027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,3584,0.010040889183680216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,16384,0.12780977620018855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,64,0.004570666700601578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,3072,0.034006221426857844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,256,0.006949333681000604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,16384,0.021171554923057556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,8192,0.7110204696655273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,10240,0.7775857713487414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,4096,0.06424088610543145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,65536,0.21996266312069365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,1024,0.007900444169839224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,64,0.0063795554969045855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,3584,0.04239644606908163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,7168,0.043897777795791626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,768,0.007544000115659501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,12288,0.08651644653744167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,2048,0.04189155499140421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,51200,2.3115306430392795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,3072,0.036397334602144026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,32,0.010812444819344414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,64,0.019017777509159513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,51200,0.012552888856993781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,2048,0.007196444604131911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,2560,0.08784621953964233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,16384,0.4260479874081082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,51200,0.1011013322406345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,1536,0.010575110713640848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,7168,0.028024888700909082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,2048,0.05383200115627713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,51200,0.9776773452758789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,8192,0.014154665999942355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,512,0.00461422238085005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,3072,0.006855999843941794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,7168,0.6854355600145129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,256,0.005174222091833751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,512,0.0033448889023727844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,16384,0.07457155651516385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,64,0.0028284444577164124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,10240,0.7869333161248101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,5120,0.1514106723997328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,3584,5.438260396321614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,12288,0.025116443634033203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,5120,0.23871734407213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,6144,0.03435466686884562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,12288,0.2216026650534736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,256,0.007641777396202087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,7168,0.19233511553870308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,16384,0.11636355188157822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,256,0.005447111195988125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,128,0.003909333298603694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,4096,0.02700977855258518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,2560,0.011506666739781698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,2048,0.0184871107339859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,3072,0.01679377754529317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,65536,0.07084621985753377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,12288,0.45572180218166775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,3072,0.012247111234400006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,1024,0.007028444773621029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,768,0.007738666401969061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,32,0.004474666797452503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,32,0.002830222249031067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,7168,0.022031999296612207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,10240,0.2782391177283393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,1536,0.01183111137813992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,32,0.006124444305896759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,8192,0.014747555057207743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,12288,0.08994844224717881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,768,0.005391110976537068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,64,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,3072,0.016505777835845947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,512,0.007667555577225155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,8192,0.10294310914145575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,7168,0.046075556013319224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,64,0.04790844519933065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,32,0.007672888537247975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,32,0.004963555683692296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,2048,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,12288,0.13160444630516901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,7168,0.11001600159539117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,1024,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,3584,0.007592000067234039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,3584,0.05217599868774414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,2560,0.007173333730962541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,1536,0.02088533341884613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,16384,0.019175110591782462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,5120,0.1365893284479777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,2560,0.05547555618815952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,4096,0.13682044876946345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,2560,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,3584,0.4693911340501573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,6144,0.007851555943489075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,256,0.006414222220579783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,64,0.0399582220448388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,3584,0.007790221936172909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,256,0.004443555656406614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,4096,0.09750133090549046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,32,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,12288,0.3413280116187201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,128,0.01145511120557785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,7168,0.013810666898886362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,64,0.004017777740955353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,3072,0.010211555494202508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,4096,0.10953599876827663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,6144,0.010080888867378235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,768,0.01349866638580958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,4096,0.013901333014170328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,1536,0.005642666584915585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,1024,0.012676444318559436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,512,0.015295111470752291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,768,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,1024,0.011334222224023608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,512,0.004914666629499859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,65536,0.06894489129384358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,2048,0.043191111750072904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,7168,0.05242400036917793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,7168,0.01640888883007897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,3072,0.01315111087428199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,51200,0.8828124470180936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,256,0.004843555390834808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,10240,0.07150133450826009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,3584,0.015536889433860779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,3072,0.06998844279183282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,768,0.13595110840267605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,512,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,512,0.03464355402522617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,3584,0.007368889119890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,2560,0.06547555658552381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,2560,0.01953066719902886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,32,0.006257777826653586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,4096,0.006782222125265334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,3584,0.03558489018016391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,32,0.021433777279324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,7168,0.013199110825856527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,128,0.0028568889117903183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,768,0.09647111097971599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,32,0.002765333279967308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,6144,0.1451760000652737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,64,0.00665244460105896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,128,0.0028408887899584244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,64,0.0026444445053736367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,10240,0.050698667764663696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,10240,1.1289040247599285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,10240,0.01959377858373854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,51200,0.01401511165830824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,8192,0.013918222652541267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,4096,0.09388177924686009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,6144,0.3692186673482259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,3072,0.009875555833180746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,256,0.0052355556852287715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,1024,0.006046222315894232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,65536,0.06724177466498481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,5120,0.01699466672208574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,51200,0.11114488707648383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,256,0.017464889420403373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,5120,0.019215111931165058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,1024,0.008845333423879411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,16384,0.01733955575360192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,51200,0.5773964458041722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,2560,0.05603910817040337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,3584,0.21850221686893037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,2048,0.34168799718221027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,768,0.005838222387764189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,7168,0.016176000237464905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,1024,0.012131555212868584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,256,0.04007555709944831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,768,0.007292444507280986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,16384,0.3556942145029704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,256,0.005662222289376789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,2048,0.04640000065167745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,2048,0.02038666605949402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,10240,0.027624888552559748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,1536,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,2560,0.018728888697094388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,3584,0.0181031111213896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,5120,0.01554844445652432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,2560,0.06922666894065009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,16384,0.18730933136410186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,512,0.005394666559166378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,512,0.0053680001033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,6144,0.13201688395606145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,2048,0.05072888731956482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,3072,0.009630222287442949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,768,0.005839111076460944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,16384,0.11524532900916205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,3584,0.08457333511776394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,7168,0.020357333951526217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,1536,0.61662843492296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,10240,0.03398488958676656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,65536,0.15486578146616617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,3072,0.029281778468026057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,1024,0.012425777812798819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,128,0.0027582223216692605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,2560,0.035564445787005954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,32,0.0030088888274298776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,10240,0.037407110134760536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,16384,0.3665911091698541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,3584,0.011362666885058085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,5120,0.22708710034688315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,512,0.004784000002675586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,128,0.005951111101441913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,2560,0.04466755522622002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,65536,3.819915347629123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,1024,0.04338399900330437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,2048,0.09625244140625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,768,0.004464000049564573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,1536,0.008746667040718926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,12288,0.17163466082678902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,8192,3.4163920084635415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,5120,0.06845955716239081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,512,0.0037840000457233856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,51200,1.08278136783176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,3072,0.06552088922924466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,64,0.015575110912322998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,128,0.013276444541083442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,512,0.0041599998043643106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,64,0.0031306667046414483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,256,0.005660444498062134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,12288,0.305299546983507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,6144,0.011590222517649332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,51200,0.5845244195726182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,10240,0.7245626449584961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,10240,0.03599377804332309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,1536,0.21378844314151338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,7168,0.035790221558676824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,16384,0.16064977645874023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,32,0.005192888693677055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,5120,0.084133333630032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,1536,0.029145777225494385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,10240,0.020483555065260995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,10240,0.0317466656366984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,3072,0.13375022676255968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,8192,0.014440889159838358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,256,0.012649777862760754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,51200,2.6598267025417752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,7168,0.4307075606452094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,768,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,32,0.018156444032986958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,128,0.0031226666437255014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,3072,0.02392622166209751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,10240,0.045250667466057666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,2048,0.11349066098531087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,7168,0.02199555602338579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,512,0.0058542220956749385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,8192,3.753501468234592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,4096,0.1865208943684896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,51200,0.13295200135972765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,64,0.016985777351591323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,6144,0.10478666755888198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,6144,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,8192,0.37289778391520184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,6144,0.05396533343527052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,256,0.0034871109657817413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,512,0.016855110724767048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,16384,0.018331555856598746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,16384,0.025241777300834656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,128,0.007723555796676212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,12288,0.019889778561062284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,7168,0.043022221989101835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,128,0.01661066710948944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,2048,0.03583466675546434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,32,0.00369422220521503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,12288,0.3429866631825765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,64,0.002987555538614591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,5120,0.2649386723836263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,512,0.010133333504199982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,512,0.00313688897424274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,4096,0.10787022113800049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,7168,0.024640000528759424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,4096,0.052849776215023465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,3584,0.008584000170230865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,10240,0.017895999881956313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,768,0.005588444570700328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,64,0.019174221489164565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,512,0.016047999262809753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,64,0.005374222166008419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,1536,0.007322666545708974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,512,0.0053013331360287136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,256,0.0030053332448005676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,7168,0.028704888290829126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,2048,0.5626364284091526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,128,0.005997333261701796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,6144,0.007094221810499827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,128,0.003338666632771492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,4096,0.011277333729796939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,256,0.013991999957296582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,2560,0.006521777974234686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,128,0.020367999871571858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,2048,0.015204444527626038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,128,0.004549333204825719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,3584,0.029088000456492107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,768,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,5120,0.030984888474146526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,3584,0.35965421464708114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,2048,0.040248887406455144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,16384,0.07557066943910387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,51200,1.6790044572618272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,4096,0.37197065353393555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,6144,0.017735110388861764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,7168,0.01866311166021559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,6144,0.012301333248615265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,768,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,2560,0.027403554982609216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,2048,0.01791733337773217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,768,0.005544888890451855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,512,0.02604177759753333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,2560,0.028301331732008193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,8192,0.22712267769707573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,256,0.24181776576571992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,2048,0.013825777504179211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,768,0.0058853332367208265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,12288,2.9115698072645397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,10240,0.5434115727742513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,128,0.004321777572234471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,5120,0.33923734558953184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,32,0.00648266656531228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,2048,0.07464622126685248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,7168,0.03144800000720554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,51200,18.36773681640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,5120,0.011549333731333414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,32,0.02171733313136631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,12288,0.27183466487460667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,12288,0.024453333682484094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,512,0.009676444033781687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,6144,0.02480711042881012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,8192,0.02979288829697503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,51200,0.38719820976257324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,64,0.0074622225430276655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,2048,0.008189332981904348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,12288,0.27678577105204266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,3072,0.007619555625650618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,32,0.002499555548032125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,12288,0.21323555045657686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,8192,0.0127688886390792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,65536,0.09276266892751057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,4096,0.11024889018800522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,65536,0.07142133182949491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,512,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,2560,0.01293511109219657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,6144,0.01831733352608151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,65536,0.0728000005086263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,6144,0.010401777923107147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,2048,0.009929777847396003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,1536,0.034436444441477455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,16384,0.023294222023752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,768,0.04238577683766683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,512,0.005069333232111401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,5120,0.006240889016124938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,4096,0.008986666798591614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,128,0.003019555575317807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,5120,0.040066666073269315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,512,0.004594666676388847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,3584,0.006746666712893381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,1536,0.030809776650534734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,512,0.0063324446479479475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,1024,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,8192,0.05804710918002658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,3584,0.01644888851377699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,2560,0.3725279966990153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,5120,0.12076444096035427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,2560,0.1383777724372016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,10240,0.16007199552324083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,512,0.01936533384852939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,256,0.04677599999639723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,2048,0.01198755535814497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,32,0.008928889201747047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,4096,0.00814666681819492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,128,0.0029235554652081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,64,0.007628444168302748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,4096,0.05124533176422119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,4096,0.012963555753231049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,16384,0.0770346654786004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,2048,0.01533066729704539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,1024,0.010911111202504901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,10240,0.2333271238538954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,2560,0.09874577654732598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,5120,0.04593511091338264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,5120,0.02418488926357693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,65536,0.08351911438835992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,16384,0.6978960037231445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,12288,0.3003404405381945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,16384,0.006139555739031897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,64,0.04108533263206482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,6144,0.30560800764295787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,768,0.03119200136926439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,16384,0.5214897791544596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,12288,0.8395911322699653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,5120,0.1131697760687934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,2048,0.006698666761318843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,12288,0.08550577693515354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,12288,0.08779199918111165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,1024,0.005664888769388199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,512,0.00371111101574368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,32,0.024118221468395654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,8192,0.01661155621210734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,5120,0.07579288880030315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,8192,0.01110755569405026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,1536,0.011054221954610614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,256,0.010766222245163388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,1024,0.03642044464747111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,256,0.008191111187140146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,2560,0.008404444489214156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,51200,0.02552977701028188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,4096,0.018336888816621568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,512,0.026561778452661302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,6144,0.12035111586252849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,128,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,5120,0.019651555352740817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,512,0.08872533506817287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,128,0.003076444483465619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,6144,0.2691448794470893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,256,0.005620444400442972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,2560,0.03035466538535224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,51200,0.09587644206153022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,16384,0.07214399841096666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,6144,0.014038221703635322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,65536,0.1858382225036621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,3584,4.586764441596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,32,0.0028213332924577924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,16384,0.020578665865792167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,1024,0.03174399998452928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,128,0.006222222414281633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,512,0.015435554915004306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,512,0.0035208890007601846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,16384,0.11350666152106391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,10240,0.07437866926193237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,3584,0.012472000386979846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,128,0.07207022110621135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,65536,1.731600867377387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,1024,2.004327138264974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,16384,0.02847822176085578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,51200,0.06577244732115003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,1536,0.01404711107412974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,1024,0.015423110789722867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,5120,0.012719111310111152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,256,0.0059057776298787855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,2560,0.008196444147162968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,2048,0.02700622214211358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,2048,0.07591999901665582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,64,0.004971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,12288,0.49537335501776797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,10240,0.12969066037072077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,6144,0.043360888957977295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,32,0.006540444576077991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,32,0.006605333338181178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,6144,0.023619555764728125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,6144,0.013707555830478668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,7168,0.0921048919359843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,7168,0.04415555463896858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,1024,0.02144355575243632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,7168,0.012907555533779992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,3584,0.04062311185730828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,65536,0.6651395691765679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,512,0.02273066673013899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,3584,0.007552000383536021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,2048,0.005480000128348668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,3072,0.0931546688079834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,12288,0.4279919995201959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,6144,2.320036358303494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,64,0.011782222323947482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,65536,0.2009777757856581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,3584,0.00703022215101454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,768,0.006040888941950268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,768,0.01835911141501533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,51200,0.057429333527882896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,12288,0.22257600890265572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,7168,0.014317333698272705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,1024,0.6006586816575792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,2048,0.006454222318198945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,5120,0.09249955415725708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,7168,0.011363555159833698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,12288,1.0785626305474174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,1536,0.27355554368760854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,1536,0.006770666688680649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,10240,0.04775289032194349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,256,0.006048000107208888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,65536,0.03162755568822225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,1536,0.007631999750932057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,16384,1.201819525824653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,5120,0.026504000027974445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,16384,0.10471289025412665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,256,0.013456888496875763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,5120,0.032568001084857516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,256,0.011094222466150919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,6144,0.009824000298976898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,10240,0.33990221553378636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,4096,0.02232088810867733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,51200,2.043438169691298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,5120,0.1203920046488444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,4096,0.013088000317414602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,7168,0.029467556211683486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,2560,0.5712381998697916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,51200,0.11914755238427056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,8192,0.04078400135040283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,2048,0.014456000592973499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,5120,0.02831377916865879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,10240,0.03487822082307603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,12288,0.015619556109110514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,4096,0.10184533066219753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,5120,0.03910933269394769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,10240,0.034918222162458636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,7168,0.2535635630289714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,5120,0.016155555844306946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,5120,0.015855999456511605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,10240,0.21689955393473306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,7168,0.032102220588260226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,7168,0.05726933479309082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,51200,0.1211857795715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,256,0.00444177786509196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,8192,0.05264977614084879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,12288,0.027423111928833857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,3584,0.009897777603732215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,7168,0.08399111032485962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,65536,2.40738402472602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,64,0.012911111116409302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,8192,0.21473956108093262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,3072,0.02221422228548262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,6144,0.01222222215599484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,6144,0.18581155935923258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,3072,0.0068764446510208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,16384,0.01819022165404426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,3072,0.012512000070677863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,768,0.0472924444410536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,32,0.0028853333658642242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,512,0.015740444262822468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,65536,0.07742755280600654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,1024,0.1471502251095242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,3584,0.13157244523366293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,8192,0.016442666451136272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,8192,0.024178667200936213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,12288,0.07586133480072021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,2560,0.009039999710188972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,64,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,65536,0.8277280065748426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,5120,0.1652631097369724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,512,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,64,0.004639111045334074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,51200,5.650007883707683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,2560,0.2430666552649604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,65536,0.08262488577100965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,65536,0.3986062208811442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,1536,0.02103999919361538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,10240,0.015987555185953777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,2048,5.259036593967013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,6144,0.013200889031092325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,2560,0.010913777682516309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,12288,0.01722666621208191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,7168,0.017039111918873258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,512,0.057358225186665855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,512,0.00462311092350218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,12288,4.771047804090712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,7168,0.13550221920013428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,4096,0.007174222005738153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,256,0.01759733259677887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,65536,0.014436443646748861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,65536,0.06798488563961454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,65536,0.07508622275458442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,16384,0.43697688314649796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,64,0.006102222121424145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,1024,0.003863111138343811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,768,0.009162666896979014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,10240,0.015678221980730694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,2048,0.03815911213556925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,10240,0.7442542182074653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,512,0.04340622160169813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,51200,1.3958035575018988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,51200,14.535987006293404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,2048,0.049597332874933876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,1536,0.009650666680600908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,5120,0.009396444592210982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,256,0.007356444166766272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,256,0.018011555075645447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,10240,0.012806221842765808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,4096,0.009315555294354757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,128,0.007921777665615082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,2048,0.04238311118549771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,8192,0.14003288745880127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,3584,0.013416888813177744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,256,0.01144088887506061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,5120,0.07115199830796984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,65536,0.5330710940890843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,2048,0.007002666592597961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,3584,0.01531555586391025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,768,0.006142222219043308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,2560,0.7103164460923937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,3072,0.008054222497675154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,32,0.0038177776667806837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,2560,0.09352355533175999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,5120,0.033208890093697443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,128,0.03379377722740173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,10240,3.600131564670139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,16384,0.21352710988786486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,768,0.0042071110672420925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,3584,0.010662222074137794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,3584,0.04607466525501675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,2560,0.25378134515550405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,4096,0.16048533386654323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,3072,0.013491555220550962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,32,0.11587288644578721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,10240,0.025938666529125635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,768,0.016964443855815463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,64,0.005277333160241445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,256,0.011225777367750803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,768,0.0173688895172543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,2560,0.03289244572321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,7168,0.0295724438296424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,3072,0.010312889185216691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,2560,0.015478221906556023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,3584,0.01843377782238854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,12288,0.272032896677653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,512,0.005294222384691238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,6144,0.8901519775390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,7168,0.056576000319586865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,3072,0.0502417782942454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,64,0.006392000036107169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,3072,0.00942311104800966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,1024,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,12288,0.016708445217874315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,256,0.004503111044565837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,6144,0.15386133723788792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,3584,0.0118222220076455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,5120,0.009768000079525841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,51200,17.86309644911024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,8192,0.03152533372243246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,2048,0.007502222226725684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,64,0.008064889245563084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,8192,0.09918044673071967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,1536,0.011135111252466837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,256,0.0028773333049482773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,64,0.0029582221888833572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,65536,0.07042933172649808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,1024,0.008617777791288164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,16384,0.22859112421671549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,12288,0.1270444393157959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,16384,0.042103999190860324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,4096,0.22951110204060873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,7168,0.02387733260790507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,512,0.0031235555393828284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,51200,0.1145448949601915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,2048,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,16384,0.5254889064364964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,4096,0.023090667194790308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,32,0.016377778516875375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,51200,0.15478132830725774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,512,0.11882755491468643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,1024,0.021729777256647747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,1024,0.005708444449636672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,16384,0.04097955425580343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,5120,0.011746666497654386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,6144,0.007747555772463481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,3584,0.02978044417169359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,2560,0.07942133479648165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,16384,1.673492431640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,5120,0.01126488877667321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,65536,0.3694435490502252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,768,0.005341333233647876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,10240,0.04207288887765673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,51200,0.2812275621626112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,5120,0.009170666337013245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,12288,0.07946133613586426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,5120,0.0192577772670322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,6144,0.05202310946252611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,32,0.003297777846455574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,3072,0.05295644534958733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,768,0.006257777826653586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,8192,0.1790462202495999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,64,0.00267999991774559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,5120,0.0214355554845598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,12288,0.12544800175560847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,128,0.01297155519326528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,12288,0.0313368903266059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,51200,4.426931593153212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,5120,0.011176889141400656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,2048,0.010689777632554373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,65536,25.753587510850693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,1536,0.007128888534175024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,2048,0.023231999741660223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,3584,0.016414221790101793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,7168,0.2099484470155504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,256,0.010410666465759277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,2560,0.007674666742483775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,256,0.002990222225586573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,768,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,32,0.10834666755464341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,1024,0.006983111302057902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,2048,0.11151644918653701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,16384,0.4627102216084798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,32,0.008603555460770925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,1536,0.010950222611427307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,8192,0.05078399843639798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,12288,0.015062222878138224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,512,0.008172444171375698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,7168,0.020106666617923312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,2048,0.040545778142081365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,1536,0.005895111295912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,3072,0.27176356315612793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,1024,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,32,0.0031644445326593188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,51200,1.2262017991807725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,512,0.09068799681133694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,8192,0.2768497731950548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,7168,0.016525333126386006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,12288,0.07299911313586764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,256,0.11302933428022598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,1024,0.02739022175470988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,16384,1.2324399948120117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,10240,0.08821244372261895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,6144,0.06519822279612224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,32,0.026851556367344324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,10240,0.01953066719902886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,768,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,5120,0.020017777880032856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,768,0.0064284445510970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,51200,0.05968799855973986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,2560,0.06246933009889391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,5120,0.01126399967405531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,128,0.007375111182530721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,10240,0.34433333079020184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,3072,0.015440000428093804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,768,0.004214222232500712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,5120,0.2782844437493218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,32,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,512,0.02334844403796726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,1536,0.016208888755904306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,128,0.005094222310516569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,12288,0.008059555457697975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,51200,0.8133111000061035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,128,0.004971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,64,0.002587555597225825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,6144,0.02611200014750163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,2048,0.012490666574902005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,128,0.004049777984619141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,1536,0.017052445146772597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,3584,0.016562667157914903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,51200,0.012090666426552666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,7168,0.05172533459133572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,8192,0.09072711070378621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,5120,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,768,0.006852444261312485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,4096,0.04161511196030511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,5120,0.019487111104859244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,12288,3.914981418185764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,8192,0.03279822071393331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,1536,0.0046986668474144405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,16384,0.04912711183230082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,1024,0.024637333220905725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,128,0.020506666766272653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,16384,0.02292355563905504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,32,0.014175110393100314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,6144,0.027126222848892212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,2048,0.009750222166379293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,5120,0.07545422183142768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,3584,0.04288533329963684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,3072,0.013413333230548434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,2048,0.026130666335423786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,6144,0.19082221719953749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,65536,0.01166133334239324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,5120,0.051157332128948636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,3584,0.022942221826977197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,16384,0.021574222379260596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,2560,0.009200000100665623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,3584,0.050550222396850586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,7168,0.05534222059779697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,512,0.016532444291644625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,2048,0.022112889422310725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,128,0.09366933504740398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,65536,0.09899022181828816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,5120,0.011491555306646558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,3584,0.00797333319981893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,1024,0.01865511139233907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,2048,0.01038933296998342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,32,0.031911111540264554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,3072,0.05580621957778931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,3072,0.014592889282438489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,10240,0.006157333238257303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,768,0.023556444380018447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,51200,0.08591644631491767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,4096,0.013577777478430005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,5120,0.009667555491129557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,512,0.016711999972661335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,64,0.0025173332542181015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,5120,0.019697777099079557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,5120,0.1810506714714898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,65536,0.06630755795372857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,2560,0.025787555509143408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,2560,0.03374400072627597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,2560,0.1825653314590454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,16384,0.07137422429190741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,5120,0.042351110113991626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,6144,0.00817599975400501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,16384,0.12184445063273112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,64,0.03196622265709771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,12288,0.4026622242397732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,8192,0.011672000090281168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,6144,0.010189333723651038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,65536,0.29205067952473956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,5120,0.16930577490064833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,6144,0.009602666729026372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,768,0.07524977789984809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,512,0.01577333278126187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,768,0.0037928890022966596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,3584,0.017989334132936265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,3584,0.015220445063379077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,64,0.0026008888251251644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,6144,0.012816888590653738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,4096,0.06338844696680705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,32,0.00869688888390859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,128,0.032815112007988825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,10240,1.1985404756334093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,512,0.008211555580298105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,4096,0.010775999890433418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,5120,0.22744978798760307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,1024,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,7168,0.012915555801656512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,2560,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,1024,0.008408889174461365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,2048,0.01867377758026123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,65536,0.1737875541051229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,51200,0.40354487631056046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,2048,0.007850666840871176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,3584,0.6667751206292046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,65536,0.026130666335423786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,1536,0.03255733185344272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,512,0.01976355579164293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,12288,0.05510488814777798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,256,0.0032319999817344877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,1536,0.009111111362775167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,7168,0.010867555936177572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,256,0.0615288880136278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,768,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,10240,12.778037177191841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,5120,0.01017777787314521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,1024,0.004910222358173794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,256,0.010087110930018956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,1024,0.02451466686195797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,16384,0.020279111133681405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,512,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,1536,0.00793599999613232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,128,0.007054222126801808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,8192,0.027822222974565294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,51200,0.052433778842290245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,1024,0.04242222176657783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,3072,0.008150222400824228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,12288,0.1857102182176378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,6144,0.10989511013031006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,2048,0.00740533322095871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,3072,0.045851555135515004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,6144,0.01293244461218516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,512,0.016734222571055096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,51200,0.11352354950375027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,256,0.004512000001139111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,3072,0.11065600315729777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,1536,0.01604977746804555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,51200,0.08043289184570312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,7168,0.13287911150190565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,51200,0.04389511214362251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,16384,0.046112000942230225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,1024,0.01126399967405531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,4096,0.01484266585773892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,16384,0.022130666507614985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,4096,0.017785777648289997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,4096,0.010461333725187512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,12288,0.031269331773122154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,64,0.005467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,4096,0.03630044394069248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,8192,0.02273955610063341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,12288,0.3206133312649197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,128,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,1536,0.00602311102880372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,1024,0.01425866617096795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,6144,0.020925333102544148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,51200,0.31379821565416127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,2560,0.35776355531480575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,256,0.013824000126785703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,7168,0.014085332552591959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,2048,0.014835556348164877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,65536,0.10837777455647786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,1536,0.016951110627916124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,128,0.10820177528593276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,32,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,32,0.0027671110712819626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,12288,0.07222133212619357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,51200,0.8743502298990885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,5120,0.025342222717073228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,2560,0.011663999822404651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,128,0.049144887261920504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,3584,0.012633778154850006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,4096,0.0063484443558586975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,1536,0.008031999899281396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,64,0.008240888516108194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,65536,0.1236906713909573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,51200,2.225205315483941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,2048,0.01387288918097814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,1536,0.008391111261314815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,65536,0.17501066790686715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,128,0.00666222224632899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,3072,1.5932417975531683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,2048,0.008723555339707268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,1536,0.015240889456537036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,32,0.011121778024567498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,1536,0.0513555539978875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,512,0.004120888809363048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,3584,0.07454222440719604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,768,0.02058844433890449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,6144,0.06362044148974948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,65536,0.1044915583398607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,256,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,2048,0.007182222273614671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,256,0.18446844153934053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,5120,0.2329795625474718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,512,0.0033075554917256036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,5120,0.25664620929294163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,64,0.008017777568764156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,128,0.027072888281610277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,7168,0.03254400028122796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,51200,0.33044089211357963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,4096,0.02004888819323646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,2048,0.013291555146376291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,512,0.0177511109246148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,16384,0.00999111102686988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,6144,0.006365333166387346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,128,0.2089137766096327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,5120,0.18204266495174834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,128,0.004414222306675381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,51200,1.3810560438368056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,8192,0.01703466640578376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,1024,0.009858667022652095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,128,0.009152000149091085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,6144,0.05072177780999077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,1024,0.33193421363830566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,16384,0.03854755560557047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,7168,0.21299378077189127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,2560,0.05111733410093519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,6144,0.021350221501456365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,4096,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,10240,0.3467546568976508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,12288,0.17281956142849395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,8192,0.31250045034620494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,65536,3.7680257161458335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,256,0.018601778480741713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,1024,0.006617777877383762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,64,0.002691555561290847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,3072,0.22389867570665148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,32,0.0053075556125905775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,64,0.002712888850106133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,4096,0.012671999633312225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,10240,0.06008888615502251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,512,0.004885333279768626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,512,0.010862222148312463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,3072,0.058006220393710665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,128,0.006568000134494569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,768,0.003671110918124517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,1024,0.004497777670621872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,768,0.023215999205907185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,3584,0.043084445926878184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,32,0.01070755554570092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,256,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,512,0.27314310603671604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,8192,0.0566142201423645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,256,0.08397422234217326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,51200,0.05523822373814053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,16384,0.10393066538704766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,1024,0.012993777791659037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,256,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,2048,0.11671555042266846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,7168,0.15459022257063124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,7168,0.015412444869677225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,512,0.013920888304710388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,2048,0.014683556225564746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,1024,0.011718221836619906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,2048,0.057769775390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,2048,0.006650666809744305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,7168,0.01995466649532318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,6144,0.023896889554129705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,3584,0.0425217780801985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,1024,0.005696889013051987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,12288,0.2138364447487725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,12288,0.1370284424887763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,3072,0.22716887791951498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,512,0.006578666468461354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,32,0.01094044413831499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,3072,0.6421253416273329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,8192,0.03919733232922024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,6144,0.055902222792307533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,16384,0.1037928859392802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,51200,1.6293680402967665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,512,0.01907199952337477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,768,0.004328888737493091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,10240,0.2612160046895345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,2560,0.009544000029563904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,16384,0.11535467041863336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,2048,0.13327377372317845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,1024,0.005524444497293896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,12288,4.916522555881076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,128,0.007250666618347168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,51200,1.6394373575846355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,256,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,10240,0.06335644589530097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,65536,0.24809511502583823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,12288,0.03651555710368686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,8192,0.014533332652515836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,3584,0.007137777904669444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,2560,0.006076444354322221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,8192,0.2089404397540622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,2048,0.009567110902733272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,6144,0.11119288868374294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,2048,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,65536,0.6174764633178711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,128,0.00444533344772127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,12288,0.015621332658661736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,128,0.004359111189842224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,768,0.0225564440091451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,16384,0.3978453212314182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,65536,0.9431084526909722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,4096,0.014984889162911309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,65536,29.96255154079861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,16384,0.03155111273129781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,768,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,768,0.008727110922336578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,1024,0.044650667243533664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,2048,0.012883555557992725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,768,0.03302222159173753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,65536,0.24849510192871094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,256,0.024548444483015273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,65536,0.07853244410620795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,5120,0.01295466638273663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,12288,0.030644443300035264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,10240,0.8139982223510742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,12288,0.009848888549539778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,7168,0.031007111072540283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,51200,3.1824141608344187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,2048,0.012797333300113678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,16384,5.382912105984158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,2560,0.05129333337148031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,7168,0.03277422322167291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,768,0.006239111224810283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,128,0.003142222141226133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,1024,0.015871110889646742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,256,0.00498844434817632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,12288,0.017887999614079792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,3072,0.15608710712856716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,512,0.0034017778105205963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,16384,0.11561511622534858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,10240,0.13664978080325657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,7168,0.07122311327192518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,7168,0.03321244319279989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,8192,0.020040889581044514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,10240,0.02292977770169576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,65536,0.2945084571838379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,512,0.015867556134859722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,1536,0.0477128889825609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,10240,0.017296888762050204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,1536,0.004652444273233414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,1536,0.06480444139904447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,65536,0.2676995595296224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,256,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,1024,0.005605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,5120,0.4827449056837294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,5120,0.016215110818545025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,12288,0.0189057770702574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,10240,0.08736533588833278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,1024,0.027686221732033625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,128,0.004110222061475118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,16384,0.29547021124098033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,3584,0.7126106686062283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,2048,0.13062577777438694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,768,0.01780177818404304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,1536,0.006484444356626934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,16384,0.11225422223409016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,512,0.010837333069907295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,51200,0.25086667802598744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,3584,0.015231110983424716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,3072,0.09317955705854629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,256,0.10065955585903591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,7168,0.1510897742377387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,10240,0.01400888959566752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,32,0.006311999840868845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,512,0.005046222358942032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,6144,0.009728000395827824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,3584,0.018516444497638278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,6144,0.012249777714411417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,5120,0.00944088896115621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,65536,0.19687022103203666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,128,0.23893422550625273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,10240,0.051810665263069995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,3072,0.18851021925608316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,3584,0.6803350978427463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,2560,0.015061333775520325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,10240,2.072337680392795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,2560,0.009185777770148383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,2560,0.01740622189309862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,51200,0.2286835511525472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,10240,0.04836000005404154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,65536,0.6726284556918674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,768,0.025478222303920325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,3584,0.7886906729804145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,4096,0.041166222757763334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,32,0.010619555910428366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,768,0.05281688769658407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,1024,0.024480889240900677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,51200,18.45138888888889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,512,0.16822844081454805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,12288,0.023111111587948267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,12288,0.15940354930029974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,512,0.010642666783597736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,512,0.03593600127432082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,512,0.02221155497762892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,512,0.006133333262470033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,10240,0.38234488169352215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,6144,0.006004444426960415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,65536,4.844441731770833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,2048,0.03257866700490316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,10240,0.041377776198916964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,2048,0.48223288853963214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,10240,0.09848533074061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,64,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,256,0.004801777915822135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,2560,0.00831822223133511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,3584,0.02462399999300639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,4096,0.08282844225565593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,32,0.004424889054563311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,12288,0.11363200346628825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,4096,0.012540444731712341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,2048,0.010225777824719747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,12288,0.028340445624457464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,256,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,768,0.04492533206939697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,12288,0.0952577789624532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,1024,0.5861306720309788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,12288,0.6701475779215494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,51200,0.1165884468290541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,7168,0.08890400330225627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,16384,0.03415288858943515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,768,0.0037804444630940757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,10240,0.016315556234783597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,3072,0.012863111164834766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,12288,0.5813431209988064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,1536,0.25133689244588214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,128,0.0052177777720822235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,5120,0.015053333507643806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,768,0.3028106689453125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,6144,0.010216000179449717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,4096,0.015027556154463025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,2048,0.007059555914666917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,512,0.03193511234389411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,4096,0.1322071154912313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,1536,0.006351111249791251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,2560,0.012279111478063794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,1024,0.014059555199411182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,4096,0.2092613379160563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,8192,0.12635733021630183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,5120,0.00923466682434082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,8192,0.3590879970126682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,3584,0.07911466889911227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,65536,0.18032799826727974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,1024,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,64,0.005497777627574072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,16384,0.503494209713406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,51200,0.03382844395107693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,2560,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,1024,0.004602666530344221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,2048,0.007837333612971837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,3584,0.12311555279625787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,768,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,1536,0.05628177854749891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,2560,0.007364444434642792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,65536,0.642680909898546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,3072,0.18597155147128633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,5120,0.02311022248533037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,128,0.004296888907750447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,6144,40.70952351888021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,256,0.006566222343179915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,512,0.007896888587209914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,1024,0.009124444590674506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,4096,0.01854044364558326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,7168,0.025057777762413025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,768,0.01979555520746443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,3584,0.015201777219772339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,512,0.005497777627574072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,768,0.004912000149488449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,4096,0.011373333632946014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,128,0.006585777633719974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,1024,0.05123111274507311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,65536,0.15047378010219997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,5120,0.019705777366956074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,65536,0.09397066964043511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,4096,0.0313937763373057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,65536,0.21336444218953451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,5120,0.07301421960194905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,2560,0.014632888966136508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,12288,0.04237688912285698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,768,0.03743377659055922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,768,0.003458666718668408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,65536,0.1823457744386461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,256,0.004173333446184794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,2560,0.010604444477293225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,4096,0.035466667678621076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,5120,0.009709333380063375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,32,0.010097777677906884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,6144,0.011729777687125735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,10240,0.1777200036578708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,8192,0.023220444718996685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,16384,0.03341955608791775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,10240,0.01440622243616316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,1536,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,512,0.013538666897349887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,2560,0.026352889007992212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,3584,0.02714755634466807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,51200,0.45857954025268555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,2048,0.08611822128295898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,4096,0.014531556102964612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,3072,0.009545777406957414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,51200,1.2056524488661025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,51200,1.017796410454644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,16384,0.1461271047592163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,1024,0.005588444570700328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,6144,0.011726222104496427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,4096,0.08034221993552314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,5120,0.16994844542609322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,4096,0.014418666561444601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,128,0.018759111563364666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,7168,0.019509333703253005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,2048,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,32,0.004975999808973736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,256,0.024454222785101995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,4096,0.12301777468787299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,128,0.003069333318206999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,512,0.005547555370463266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,64,0.002875555513633622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,16384,0.04096622268358866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,7168,0.03860622313287523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,51200,0.056976000467936196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,65536,2.529404534233941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,10240,0.12553511725531682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,7168,0.17823378245035806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,51200,30.618642171223957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,6144,0.057822220855289035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,512,0.010185778141021729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,64,0.00814488861295912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,512,0.013361777696344586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,10240,0.08669510814878677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,3072,1.4525057474772136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,512,0.005588444570700328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,4096,0.1821368932723999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,768,0.007594666547245449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,1024,0.02991644541422526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,4096,0.010456889039940305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,3072,0.12273689111073811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,2048,0.01791111131509145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,10240,0.013592888911565145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,1024,0.004060444318585926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,512,0.019741333193249173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,2560,0.015085332923465304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,12288,0.019117333822780185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,7168,0.24842400021023223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,3584,0.015207999282413058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,256,0.00518488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,2048,0.02388622197839949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,65536,0.24743111928304037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,3584,0.010267555713653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,65536,0.14602577686309814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,51200,0.2312364313337538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,128,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,12288,0.02648266653219859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,51200,0.10618844297197129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,3584,0.011517333487669626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,16384,0.027303111222055223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,2560,0.00794044468137953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,128,0.004601777841647466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,4096,0.09053777986102635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,64,0.00683822234471639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,16384,0.13918577300177679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,12288,0.181478222211202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,2560,0.016164445214801364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,1024,0.00979644474056032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,5120,0.011219555305110084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,2560,0.01864177816443973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,65536,0.06572888957129584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,3584,0.0770053333706326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,8192,0.05323466658592224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,768,0.0036657779581016968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,10240,0.09346489111582439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,64,0.005609777652555042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,2560,0.0432533323764801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,1024,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,3584,0.04508799976772732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,16384,0.16292711098988852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,1536,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,7168,0.14367822806040445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,16384,0.07247733407550387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,512,0.004484444442722532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,7168,0.012960889273219638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,2048,0.0687697794702318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,4096,1.1808320151435006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,6144,0.12217511071099176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,2560,0.011942221886581846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,4096,0.04180533356136746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,12288,0.2963004377153185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,12288,0.1001288890838623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,32,0.0029333333174387612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,65536,0.5827875667148167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,10240,0.31713689698113334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,512,0.010269333091047075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,65536,0.21522488858964708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,256,0.0029111111329661477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,16384,0.37286310725741917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,768,0.006294222341643439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,3584,0.01220266686545478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,8192,0.013452444639470843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,16384,1.413279109530979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,6144,0.02418222195572323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,3584,0.020473778247833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,1024,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,5120,0.01773066653145684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,32,0.003064888839920362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,512,0.2466853459676107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,3072,0.02490844494766659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,2048,0.006631999793979857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,51200,0.1523324383629693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,7168,0.0428017775217692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,6144,0.8433635499742297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,768,0.04790311058362325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,2560,0.016970665918456185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,5120,0.023412444525294836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,3584,1.2048977745903862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,3072,0.41750134362114805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,1536,0.03169155451986525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,51200,0.4331324365403917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,8192,0.09713244438171387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,7168,0.006558222075303395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,3584,0.047750221358405225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,256,0.010751999914646149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,16384,0.3859528965420193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,16384,0.022071111533376906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,8192,0.12441778182983398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,8192,0.011701333026091257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,32,0.00282577777074443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,65536,0.18480000231001112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,2048,0.4228524367014567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,2560,0.03612444467014737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,2048,0.05048977666431003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,7168,0.0341306659910414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,12288,0.3541253407796224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,10240,0.019485332899623446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,10240,3.1601617601182728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,32,0.0035617777870761026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,512,0.004908444566859139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,3584,0.009043555292818282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,2560,0.006704888823959563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,8192,0.031049778064092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,128,0.014155555102560254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,512,0.011041777829329172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,5120,0.1071431106991238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,4096,0.007802666889296637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,5120,0.014807111687130399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,16384,0.25981065962049693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,1024,0.007862221863534715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,768,0.005579555614127054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,6144,0.05403822329309252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,256,0.015754666593339708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,16384,0.17031999429066977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,1536,0.005701333284378052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,768,0.04488089018397861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,128,0.005535111245181825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,12288,0.25858667161729604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,51200,0.25680443975660533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,6144,0.2593324449327257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,1024,0.016357334123717416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,4096,0.08260355393091838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,5120,0.14368889066908094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,7168,0.031719999180899724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,768,0.005587555468082428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,12288,0.02088977727625105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,2560,0.04190311166975233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,2560,1.0178693135579426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,51200,0.1194595495859782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,10240,0.019800888167487252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,1536,0.02809244394302368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,1536,0.09869066874186198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,16384,0.10293333397971259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,1536,0.024604444702466328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,6144,0.02718844513098399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,12288,0.33234667778015137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,5120,0.046837333175871104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,10240,0.02064622276359134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,2048,0.011387555963463254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,768,0.003923555629120933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,2048,0.023944889505704243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,51200,0.32770933045281303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,512,0.007672888537247975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,2560,0.008649778034951951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,256,0.01740444368786282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,65536,0.10303378105163574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,256,0.003387555480003357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,3584,0.38434576988220215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,5120,1.184073766072591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,5120,0.0631964471605089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,4096,0.014365333649847241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,32,0.0073724447025193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,3584,0.06271821922726102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,1536,0.010402667025725046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,256,0.0052764444715446895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,6144,0.3273013432820638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,1536,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,6144,0.14868888590070936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,16384,0.019998222589492798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,8192,0.03994044330384996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,16384,5.049805535210504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,5120,0.021319111188252766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,2048,0.04489422175619337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,16384,1.9940594567192926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,16384,0.02753244505988227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,10240,0.013697778185208639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,10240,0.2653262350294325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,7168,2.5946523878309464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,1536,0.008797333472304875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,2560,0.011776000261306763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,128,0.004460444466935264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,10240,2.328422122531467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,1536,0.006983111302057902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,4096,0.1348302231894599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,768,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,3584,0.6764773262871636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,2560,0.013018666870064206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,3584,0.06514310836791992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,12288,1.2349644766913521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,16384,0.05764089028040568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,12288,0.1015662219789293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,256,0.013558222187889947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,2560,0.022666666242811415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,2048,0.0069208890199661255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,4096,0.013717333475748697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,3072,0.1144711176554362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,1024,0.030792001220915053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,12288,0.5357733302646214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,3072,0.1966853274239434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,1024,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,768,0.038210666841930814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,2560,0.02344088918632931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,51200,14.340833875868055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,3072,0.017827555537223816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,2048,0.019023999571800232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,768,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,16384,0.5398133595784506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,768,0.17734666665395102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,3584,0.12694488631354436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,128,0.003724444243643019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,512,0.006576889091067844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,128,0.002977777686383989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,51200,0.7117066913180881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,16384,0.0669635534286499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,1536,0.0140684445699056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,6144,0.10047199991014268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,128,0.007448888487286038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,16384,0.029334220621320937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,6144,0.022985777921146814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,256,0.024040000306235418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,4096,0.025281776984532673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,51200,0.3617493311564128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,128,0.005417777846256892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,3072,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,3584,0.012122666670216454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,16384,0.02645155621899499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,8192,0.22459199693467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,8192,0.04786755641301473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,256,0.05773955583572388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,16384,0.3005182213253445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,3584,0.07144622008005778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,2048,0.01702044407526652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,6144,0.06723555591371325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,256,0.01848266687658098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,7168,0.014819555812411837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,3072,0.028471999698215063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,3584,0.01891644464598762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,12288,0.011140444212489657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,2048,0.015771556231710646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,8192,0.01585422290696038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,16384,0.09080800082948472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,7168,0.03570133447647095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,8192,0.013814222481515674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,2560,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,1536,0.004386666748258802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,1536,0.014808000789748298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,512,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,6144,0.024616888827747766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,512,0.0037582222786214617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,4096,0.3182951079474555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,256,0.007520889242490132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,128,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,16384,0.3949591053856744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,256,0.005708444449636672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,7168,0.02108088963561588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,3584,0.006601777755551868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,2560,0.016026667422718473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,16384,0.5132284694247776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,12288,0.39229777124192977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,64,0.003175111073586676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,51200,2.1249858008490667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,51200,0.9188551372951932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,1024,0.03218400147226121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,10240,0.4176746739281549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,3072,0.0396071109506819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,128,0.004110222061475118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,8192,0.1662586662504408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,65536,0.534162680308024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,16384,0.03762400150299072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,7168,0.02017688916789161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,2048,0.0195768889453676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,256,0.044103112485673696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,5120,0.04227200150489807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,65536,0.007956444389290279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,32,0.00272444449365139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,2048,0.007072000039948358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,128,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,10240,0.214556442366706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,12288,0.019188443819681805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,65536,0.10268355740441217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,768,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,10240,0.1643644438849555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,7168,0.04494222336345249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,51200,0.7522737714979383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,2560,0.008330666356616551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,2048,0.011223999990357293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,16384,0.11125777827368842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,1024,0.006866666591829724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,65536,0.07326844665739271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,2560,0.03074044320318434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,512,0.03966755668322245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,64,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,512,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,7168,0.04566488994492424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,2048,0.007932444413503012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,64,0.0057964444988303715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,7168,0.4375786781311035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,10240,0.11676177713606094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,3584,0.010655110908879174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,4096,0.029888000753190782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,512,0.006092444476154115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,12288,0.021120000216695998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,8192,0.015487111277050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,5120,0.012517333030700684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,12288,0.03386133246951633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,32,0.028580443726645574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,5120,0.16940977838304308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,32,0.004982222285535601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,32,0.007888000044557784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,51200,0.27352621820237905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,768,0.06206578016281128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,1536,0.132351107067532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,32,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,12288,0.4624808629353841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,3584,0.008628444539176093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,5120,0.5406933360629612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,8192,0.008922666311264038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,6144,0.013432888521088494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,8192,0.023077333966890972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,256,0.003944888710975647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,6144,0.008175111479229396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,65536,0.012906666431162091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,51200,0.9150560167100695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,6144,0.011052444577217102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,16384,0.14663733376397028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,51200,0.14004533820682102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,16384,0.028062221076753404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,5120,0.03970311085383097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,64,0.002571555475393931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,65536,0.26662755012512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,1536,0.007442666424645319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,256,0.005183111048407025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,1024,0.06146577994028727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,768,0.02257688840230306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,65536,4.390021430121528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,2048,0.010601777997281818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,4096,0.02756711178355747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,65536,0.5642826822068956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,12288,0.011904000408119626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,10240,0.021019554800457422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,7168,0.20702044169108072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,3584,0.05743200249142117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,3072,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,128,0.005499555418888728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,16384,45.16265190972222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,12288,0.11237777603997125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,3584,1.430259598626031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,65536,2.7986000908745656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,65536,3.2079743279351125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,3584,0.07069333394368489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,7168,0.0816719995604621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,2560,0.009203555683294931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,1536,0.01884533299340142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,16384,0.019319999549123976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,64,0.01570933394961887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,3584,0.017038222816255357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,7168,1.2633448706732857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,6144,0.009381333159075843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,8192,0.15015110704633924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,768,0.005716444303592046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,3584,0.01110311100880305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,10240,0.13095822599199083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,8192,0.8563013076782227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,7168,0.6404053370157877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,51200,0.43922845522562665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,1024,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,2560,0.009406222237481011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,5120,0.016190222567982145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,768,0.003879111260175705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,32,0.006599111275540457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,3072,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,3584,0.2885857688056098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,64,0.016992888516849942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,2048,0.24727021323310006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,8192,0.028639998700883653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,256,0.003133333391613431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,768,0.00582755563987626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,51200,1.794144100613064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,64,0.005682666682534748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,51200,0.8773137728373209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,2560,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,256,0.003685333248641756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,768,0.005568888866239124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,64,0.002983111060327954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,51200,0.625673770904541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,8192,0.028359999259312946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,5120,0.024357333779335022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,6144,0.4312737782796224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,2560,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,2048,0.03547999925083584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,32,0.0029599999801980127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,512,0.025636444489161175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,6144,0.22895465956793892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,1024,0.13132977485656738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,1536,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,65536,39.64783393012153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,3584,0.14746756023830837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,5120,0.011321778098742167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,1536,0.007820444802443186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,512,0.01998933321899838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,768,0.024559110403060913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,3072,0.7449057896931967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,5120,0.021741333934995864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,2560,0.27894579039679634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,128,0.007496889266702864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,3584,0.9349848429361979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,65536,1.01365417904324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,10240,0.2456737889183892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,3072,0.009367110828558603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,51200,0.07204533285564847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,1536,0.011184888581434885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,1536,0.01683111157682207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,16384,0.5981618033515083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,2048,0.032031112247043185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,2560,0.008647110727098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,6144,4.180562761094835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,768,0.02085599965519375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,768,0.0063822223908371395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,6144,0.12167289521959092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,2560,0.07226488987604777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,2560,0.09578488932715522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,2560,0.006258666515350342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,128,0.003042666655447748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,1536,0.013471110827393003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,3584,0.020963556236690942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,256,0.004869333157936732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,65536,3.3609483506944446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,768,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,2048,0.009097778134875828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,1536,0.17340532938639322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,1024,0.004552888787455029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,5120,0.03899022274547153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,6144,0.021067554752031963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,3072,0.02330844435426924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,2560,0.007408888803588019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,3584,0.017438222964604694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,64,0.018259555101394653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,65536,0.5306062168545193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,65536,1.5325973298814561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,65536,0.23267467816670737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,5120,0.02638133366902669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,512,0.0032115555885765287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,16384,0.05521777934498257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,1024,0.04604711135228475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,4096,1.7630053626166449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,6144,0.00850844466023975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,1024,0.1336017714606391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,4096,3.5886878967285156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,51200,0.225183990266588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,64,0.005397333453098933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,51200,0.03749333487616645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,128,0.005176888985766305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,128,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,8192,0.025048888391918604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,8192,0.008760000268618265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,3584,0.010847110715177325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,65536,3.1053795284695096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,32,0.00812000036239624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,1024,0.00573244442542394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,65536,0.43468888600667316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,256,0.0058551111982928375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,1536,0.008438222110271454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,10240,0.12690756056043836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,1024,0.003969777789380816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,51200,17.891048855251736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,128,0.008016889293988546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,128,0.004484444442722532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,1024,0.05488622188568115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,128,0.004607999904288186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,10240,0.4071093400319417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,2560,0.48864179187350804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,128,0.0028773333049482773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,3584,0.019684443871180218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,6144,0.018852444158660043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,2560,0.01919911139541202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,65536,2.5543840196397567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,64,0.04725689027044508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,2048,0.02081155611409081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,32,0.002736888825893402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,1536,0.004937777916590373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,256,0.0053093334039052325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,2048,0.0064560001095136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,128,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,768,0.010278222461541494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,4096,0.11235822571648492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,256,0.008665777742862701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,128,0.004606222112973531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,512,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,768,0.005249777601824866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,51200,0.4284204377068414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,256,0.018375110295083787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,16384,0.8130177921719022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,2048,0.015429332852363586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,512,0.022341334157519873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,2560,0.0380604432688819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,768,0.007836444510353936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,7168,0.006962666908899943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,64,0.021645334031846788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,3584,0.014946666028764514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,5120,0.01740800009833442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,8192,0.06410666969087389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,1024,0.021032000581423443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,4096,2.019482718573676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,128,0.010286221901575724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,512,0.014719110396173267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,16384,0.008208889100286696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,10240,0.07225777705510457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,5120,0.04202666547563341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,51200,0.06413422028223674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,12288,2.9015714857313366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,5120,0.068904889954461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,65536,0.08470400174458821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,4096,1.6676826477050781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,65536,0.4864177703857422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,6144,0.013122667041089801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,1536,0.04695999953481886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,65536,0.6584373580084907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,16384,0.7080737749735514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,4096,0.06415377722846137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,3072,0.010890666809346942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,4096,0.009159111314349705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,65536,0.10677066776487563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,65536,5.706349266899957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,128,0.006237333433495627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,7168,0.014191110928853353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,65536,1.3505129284328883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,12288,0.0754986670282152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,3584,0.12579200002882215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,8192,0.5865493350558811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,65536,0.5141013463338217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,1024,0.12893511189354792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,7168,12.015474955240885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,2560,0.01905066602759891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,1024,0.014118222726715935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,2560,0.019652444455358718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,8192,0.008295999632941352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,1024,0.08610844612121582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,8192,0.013271110753218332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,1024,0.014691554837756686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,65536,1.0462097591824002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,2048,0.04138933287726508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,10240,0.023974221613672044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,3584,0.007730666961934831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,128,0.008917333351241218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,2048,0.0052737775776121355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,4096,0.03410044312477112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,5120,0.0064497776329517365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,1536,0.04651733239491781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,2048,0.007069333559936947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,65536,0.11770399411519368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,1536,0.016561778055297006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,2560,0.05073688758744133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,1024,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,512,0.003351111171974076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,10240,0.008771555291281806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,2560,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,10240,0.05638133154975044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,65536,0.2543297873602973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,768,0.03499733408292135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,1536,0.0053688887920644544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,8192,1.3379369311862523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,51200,1.5838328467475042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,1536,0.005311111195219888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,65536,0.3979289001888699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,16384,0.03112444281578064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,256,0.004675555560323927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,768,0.0225911107328203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,768,0.0035253332720862497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,768,0.006108444597986009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,64,0.002493333278430833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,3072,0.00904799997806549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,6144,0.05059022373623318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,4096,0.008159110943476358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,2560,0.08423733048968846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,2560,0.007347555624114142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,3072,0.03141155507829454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,256,0.0031191110610961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,51200,2.643872790866428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,7168,0.2974453237321642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,16384,0.04260266820589701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,1536,0.00656444455186526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,51200,0.647081798977322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,128,0.004010666575696734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,2048,0.03417777683999803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,2560,0.030202666918436687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,64,0.007334222396214803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,3584,0.029723554849624634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,32,0.006049777898523543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,1536,0.08339111010233562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,256,0.005120000077618493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,16384,0.020789333515697055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,1024,0.004990222139490975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,1024,0.005487111293607288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,7168,0.03794755538304647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,65536,0.07692178090413411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,10240,0.01366311146153344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,2560,0.019918221566412184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,64,0.005492444253630108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,2048,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,1536,0.3705173333485921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,32,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,51200,0.09346755345662434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,32,0.0026693333768182327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,32,0.03648089038001166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,512,0.006741333338949416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,16384,0.3137884404924181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,3072,0.06470577584372626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,7168,0.022767111659049988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,512,0.007970666719807519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,65536,1.1177742216322157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,6144,0.013764444324705334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,3072,0.028796444336573284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,1024,0.007453333172533248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,768,0.035227557023366295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,128,0.003051555612021022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,3072,0.02086044516828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,10240,0.18943466080559623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,10240,0.493785752190484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,65536,20.03526136610243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,256,0.028459555572933618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,5120,0.1736311117808024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,6144,0.008775999976529015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,5120,0.014928888943460254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,65536,0.4479822052849664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,512,0.016925333274735343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,8192,0.0534746679994795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,1024,0.0728435582584805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,3584,0.03072533342573378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,2560,0.21248533990648058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,1536,0.007831110722488826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,65536,0.7755040062798394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,4096,0.020582222276263766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,128,0.00319377767542998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,128,0.09268444114261204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,512,0.014053333136770459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,1536,0.007291555404663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,128,0.0053342220683892565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,256,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,256,0.0032106666929192017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,256,0.0069013333155049225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,65536,0.2109395530488756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,3072,0.015292444162898593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,1536,0.012873777912722694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,3584,0.014136888914638095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,256,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,3072,0.015489778584904142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,3072,0.016732444365819294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,3072,0.006430222342411677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,3584,1.0687298244900174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,2048,0.06579466660817464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,5120,0.022103110949198406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,2048,0.08354666497972275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,512,0.0035333335399627686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,3584,0.011765333513418833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,4096,0.18634933895534941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,512,0.007679999702506595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,16384,0.02013777858681149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,16384,0.023414222730530634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,2048,0.11958489153120254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,3072,0.03050666716363695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,1024,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,3072,0.009945777555306753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,10240,0.006725333217117522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,10240,0.016792888442675274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,6144,2.0291466183132596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,8192,0.017539555827776592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,4096,0.010681777364677854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,6144,0.016172443826993305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,65536,0.4223528967963324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,1536,0.007624000310897827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,6144,0.018992889258596633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,1536,0.09084622065226237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,65536,0.21240711212158203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,2560,0.006034666465388403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,2048,0.032412442896101214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,3072,0.01630577776167128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,128,0.002846222163902389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,7168,0.06801955567465888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,51200,1.0247235827975802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,51200,0.0849413341946072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,256,0.0048177776237328844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,64,0.006695999867386288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,128,0.019727110862731934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,256,0.006149333384301927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,256,0.0052791109515561005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,65536,0.23503022723727754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,65536,1.5210000144110785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,256,0.006833777659469181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,2560,0.019701333509551156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,256,0.005192000004980299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,7168,0.11703377299838597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,1536,0.03895733422703213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,65536,0.009211555951171452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,65536,0.12132355901930068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,1024,0.003882666842805015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,256,0.11650399367014568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,128,0.00489155575633049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,128,0.006863111009200414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,2560,0.016943999462657504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,12288,0.019391111201710172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,10240,0.14516622490353054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,7168,0.009909333454238044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,7168,0.06118488974041409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,1536,0.004924444274769889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,16384,0.20860532919565836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,65536,0.23066933949788412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,7168,0.054401778512530856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,32,0.005541333307822545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,51200,3.493804507785373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,5120,0.05235822333229912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,4096,1.3829697502983942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,512,0.00755555596616533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,5120,0.013306666579511432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,256,0.004600000050332812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,7168,0.01131199962562985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,512,0.0059573331640826324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,1536,0.015060444672902426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,1024,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,32,0.005221333354711533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,512,0.01182666669289271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,5120,0.4179848829905192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,1536,0.23390844133165148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,5120,0.011249778171380361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,6144,0.0058462222417195635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,65536,20.57460700141059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,1024,0.010567111273606619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,4096,0.010220444036854638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,8192,0.016516443755891588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,256,0.0191600008143319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,51200,0.5597653388977051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,51200,0.8063164287143283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,6144,0.019868445065286424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,32,0.0038453332251972626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,512,0.01726222203837501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,768,0.009008888569143083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,7168,0.012663111090660095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,32,0.002587555597225825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,256,0.011344888971911537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,16384,0.02725511127048069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,768,0.04229155513975355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,128,0.00829777783817715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,256,0.003407111184464561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,10240,0.0824728873040941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,1024,0.06986488898595174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,768,0.007727111379305522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,2560,0.08939822514851888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,2560,0.0243520008193122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,7168,0.30780087576972115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,128,0.04171466827392578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,8192,2.9710239834255643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,128,0.1072746647728814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,3072,0.011890666352377998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,12288,0.18192977375454375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,3072,0.04285866684383816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,64,0.24782310591803658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,5120,0.02327999969323476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,256,0.0028284444577164124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,4096,0.008175111479229396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,16384,0.27325601047939724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,10240,0.017655111021465726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,65536,1.4170053270128038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,8192,0.4532613224453396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,128,0.0047128887640105355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,1536,0.2366702291700575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,512,0.0033333332588275275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,1536,0.0064151109092765385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,6144,0.015139554937680563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,256,0.008346666892369589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,3584,0.011630222201347351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,64,0.0028764444092909494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,51200,0.14808711740705702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,3072,0.4957466655307346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,3072,0.0298942228158315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,1536,0.03831733266512553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,10240,4.222034878200955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,16384,8.297438727484808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,1024,0.1284773349761963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,2048,0.03428444597456191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,5120,0.14783022138807508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,64,0.016396444704797532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,128,0.0075902218619982404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,6144,1.8280497656928167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,10240,0.15154400136735705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,12288,0.03129422333505418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,10240,0.7302026748657227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,1024,0.017887999614079792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,32,0.003925333420435588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,256,0.008014221986134848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,5120,0.012159999873903064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,256,0.18185777134365508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,6144,0.008786666724416945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,1024,0.006931555353932911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,3584,0.18189422289530435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,128,0.0046675557063685525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,65536,23.51776123046875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,3072,0.20587733056810167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,32,0.018391999933454726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,6144,0.06434400214089288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,2048,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,768,0.03050044510099623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,5120,0.019631110959582858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,3584,0.018515555395020377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,768,0.010135999984211391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,3584,0.015557333827018738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,1536,0.009152000149091085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,2048,0.007746666669845581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,7168,0.027134221461084154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,12288,0.01720177796151903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,16384,0.2521573437584771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,3072,0.014171555638313293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,3072,0.011222222612963783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,3584,0.012667555775907306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,12288,0.09562844700283474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,2048,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,32,0.002862222285734283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,64,0.004918222212129169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,7168,0.011928000383906893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,7168,1.1866746478610568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,2048,0.014160000615649752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,6144,0.24519644843207467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,3584,0.013447999954223633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,16384,0.02588711182276408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,64,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,64,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,6144,0.054879112376107105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,512,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,32,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,8192,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,7168,0.027128888501061335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,51200,1.1244381798638237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,5120,0.014879999889267815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,1024,0.011054221954610614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,16384,0.07499377595053779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,16384,0.034704890516069196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,51200,1.3958444595336914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,768,0.027991112735536363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,10240,0.03613866700066461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,12288,0.02509422269132402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,65536,31.142337375217014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,3072,0.017987555927700467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,2560,0.048170665899912514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,1024,0.031166222360399034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,512,0.0034933334423436057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,12288,0.15163555410173205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,768,0.005626666463083691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,768,0.048898667097091675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,5120,0.01292711082432005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,32,0.00360977773865064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,32,0.005156444592608346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,128,0.0027422221998373666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,512,0.01093066649304496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,3072,0.036892443895339966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,5120,0.035700443718168474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,1024,0.004933333231343163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,128,0.1583120028177897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,3584,0.08790133396784465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,2560,0.09980889161427815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,3584,0.01726222203837501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,8192,0.019818666908476088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,768,0.004822222308980094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,32,0.0052693333062860705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,16384,0.21491644117567274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,3584,0.021270222134060327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,2560,0.009415999882751042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,128,0.011233777635627322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,4096,0.008243555823961893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,4096,0.0723866687880622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,12288,0.055698666307661265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,2560,0.03563822309176127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,128,0.004937777916590373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,10240,0.9884808858235677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,3072,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,2048,0.021106666988796655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,12288,0.037907554043663874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,1024,0.036425779263178505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,16384,0.050275554259618126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,512,0.005000888887378904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,256,0.0029333333174387612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,3072,0.040355556541019015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,1536,0.005802666561471091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,4096,0.008400888906584846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,16384,2.030654271443685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,4096,1.0323858261108398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,51200,0.9158684412638346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,10240,0.28375289175245494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,4096,0.8857768906487359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,12288,0.01924444403913286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,4096,0.016789333687888253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,51200,2.512358135647244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,8192,0.014912888407707214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,512,0.01146399974822998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,10240,0.5687564214070638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,768,0.0058871110280354815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,1024,0.015965332587560017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,512,0.004202666795916027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,4096,0.0147733340660731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,1024,0.0671022203233507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,7168,26.413201226128475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,1024,0.009762666291660732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,128,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,65536,0.2658310996161567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,64,0.008575999902354347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,5120,0.021327111456129286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,10240,0.03162044286727905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,1536,0.01016266644001007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,6144,0.01827555563714769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,10240,0.04294133186340332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,16384,1.402417818705241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,256,0.003674666500753827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,8192,19.380839029947918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,2048,0.02908710969818963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,3584,0.3948506779140896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,8192,0.10554133521185981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,10240,0.11999111705356175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,3072,0.030215998490651447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,3072,0.10662844445970322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,5120,0.07426844702826606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,256,0.01753866672515869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,7168,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,256,0.003342222215400802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,256,0.012481778032249875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,128,0.0035004446076022256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,1536,0.014927999840842353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,3584,0.05829866727193197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,5120,0.014825777875052558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,2560,0.0075137780772315125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,8192,0.08810310893588597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,1536,0.04348533352216085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,1536,0.04249866803487142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,1024,0.003936888857020272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,256,0.007663110891977946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,768,0.010640888578361936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,6144,0.03195910983615451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,65536,0.5026008817884657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,3584,0.02698933415942722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,4096,0.022563555174403723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,256,0.010488000180986194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,7168,0.013155555559529198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,2560,0.1722737815645006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,3072,0.40578399764166934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,3584,0.19672800434960258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,3072,0.05921955241097344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,1024,0.005108444641033809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,128,0.003346666693687439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,3072,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,10240,0.7748124334547254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,8192,0.009444444543785518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,64,0.009576888548003303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,5120,0.03142133355140686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,768,0.06975289185841878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,32,0.0094035557574696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,7168,0.09691377480824788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,8192,0.18527200486924914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,16384,0.018851555056042142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,10240,0.22161511580149332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,768,0.05713422430886162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,6144,0.4630684322781033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,6144,1.6223902172512477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,128,0.0038604442444112566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,4096,0.018283555905024212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,3072,0.013519110778967539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,128,0.0027377777215507296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,2560,0.22448799345228407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,65536,0.07032710976070829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,3584,0.1322382158703274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,2048,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,2560,2.276519987318251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,3584,0.08457599745856391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,32,0.0031351111829280853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,7168,0.02712000078625149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,16384,0.14804711606767443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,5120,0.02356622285313076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,1024,0.005868444426192178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,10240,0.07956177658504911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,7168,0.3682240115271674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,10240,0.01480888823668162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,12288,0.22228977415296766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,256,0.00552977787123786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,64,0.006538666784763336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,12288,0.020801777640978496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,1024,0.01274311128589842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,2560,0.08858311176300049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,51200,0.7064648734198676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,65536,0.7263555526733398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,64,0.010109333528412713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,3584,0.016210666961140104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,7168,0.04904177784919739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,6144,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,7168,0.04621600111325582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,768,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,64,0.0026186667382717133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,2560,0.014143110977278816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,16384,0.0306640002462599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,8192,0.05997777647442288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,5120,0.011891555454995898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,10240,0.013786666923099093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,2560,0.007920888562997183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,768,0.005208888815508948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,65536,0.2224666542477078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,1536,0.01239822225438224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,51200,0.8761128849453397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,7168,0.26990755399068195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,4096,0.06379199690288968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,1024,0.01204355557759603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,16384,2.033504909939236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,1024,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,5120,0.018332444959216647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,64,0.10660355620914036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,768,0.003995555556482739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,4096,0.012702222499582501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,512,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,2560,0.08284711175494723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,128,0.005928888916969299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,4096,0.006908444480763541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,10240,0.0203813330994712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,10240,0.020640888147883944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,512,0.0047564444442590075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,7168,0.10954399903615315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,512,0.0053395554423332214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,3072,0.015061333775520325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,64,0.010072888599501716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,3584,0.020624000165197585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,6144,0.17440356148613823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,65536,1.079427507188585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,2560,0.012607999973826937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,4096,0.30233777893914116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,768,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,512,0.006232000059551663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,7168,0.19048710664113364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,8192,0.017847999930381775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,16384,5.602249569363064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,3072,0.012506666282812754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,5120,0.46904887093438047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,4096,0.0070257774657673305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,7168,0.030777778890397813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,4096,0.0119982221060329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,2560,0.01752800080511305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,2560,0.01387466655837165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,3072,0.01591999994383918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,256,0.0038480001191298165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,3072,0.0070675553547011475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,4096,0.014458666245142618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,1536,0.00815644446346495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,8192,0.05772088633643257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,64,0.02877955635388692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,12288,0.01941777765750885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,3584,0.010077333284748925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,3072,0.010277333358923594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,51200,0.42520443598429364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,768,0.061358220047420926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,64,0.0033048888047536216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,7168,0.03985244366857741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,2048,0.007269333634111617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,4096,0.024663110574086506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,128,0.009164444274372524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,65536,0.12509332762824163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,512,0.0137155552705129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,51200,0.11953688992394342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,768,0.013401777380042605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,128,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,16384,0.0640231106016371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,51200,0.08457599745856391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,6144,0.5943751335144043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,128,0.004303110970391167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,1536,0.006968888971540663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,6144,0.042391998900307544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,128,0.002982222164670626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,10240,0.011933333343929715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,2560,0.008916444248623317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,512,0.010959111154079437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,3584,0.07287644677691989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,7168,0.009171555439631144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,10240,0.06504178047180176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,3584,0.03480622172355652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,8192,0.036510222487979464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,8192,0.4540861977471246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,65536,0.5312248865763346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,3584,0.007858666280905405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,7168,0.028239111105600994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,6144,0.12811377313401964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,12288,0.39134399096171063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,3072,0.09652711285485162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,16384,0.03918577896224128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,2560,0.010645333263609143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,3584,0.01885422236389584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,16384,0.01387555566098955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,1024,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,7168,0.0314684444003635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,768,0.029290666182835896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,16384,0.050123555792702564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,10240,0.02536444365978241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,51200,6.820725335015191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,2560,0.009920000202125972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,3072,0.026336888472239178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,51200,47.42841593424479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,8192,0.24354312154981825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,51200,0.5042995346917046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,4096,0.04457777738571167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,128,0.005642666584915585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,51200,0.3301164574093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,51200,3.996001773410373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,128,0.022230222821235657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,51200,0.1531075504091051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,8192,0.02166755497455597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,5120,0.6126124593946669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,3072,0.02071555621094174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,2560,0.08881600035561456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,12288,0.3223022090064155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,128,0.00629066675901413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,256,0.007037333316273159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,256,0.013056000073750814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,4096,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,256,0.005209777918126848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,8192,0.03154933452606201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,12288,0.06904355684916179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,51200,0.585693359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,65536,1.136416859096951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,256,0.007376000285148621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,10240,0.02163911031352149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,65536,3.138715532090929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,6144,0.0169831116994222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,2048,0.08606933222876655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,64,0.015400000744395785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,1536,0.004670222186379963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,1024,0.003744889050722122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,512,0.005407111098368962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,32,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,5120,0.008808888494968414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,51200,0.059119999408721924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,1536,0.11587644947899713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,16384,0.5225013097127279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,1536,0.007186666958861881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,7168,0.12067822615305583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,65536,0.16857511467403838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,1536,0.006664888726340399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,6144,0.012561777399645912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,10240,0.016043555405404832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,7168,0.3813226752811008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,3584,0.1133342186609904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,4096,0.008832889298597971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,7168,0.10837244325213963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,512,0.04033244318432278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,8192,0.04180711176660326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,2048,0.043015112479527794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,2560,0.03442400031619602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,512,0.021182222498787776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,512,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,3072,0.5944915347629124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,1536,0.005224888937340842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,512,0.020452444752057392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,3072,0.013193777865833707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,256,0.006976000136799282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,3584,0.019080888893869188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,32,0.009418666362762451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,1536,0.004368888835112254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,6144,1.0728382534450955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,16384,7.107420603434245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,12288,0.02687377731005351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,8192,0.014057776994175382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,6144,0.3670835494995117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,2048,0.11218844519721137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,1536,0.10836622450086807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,512,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,65536,6.490821414523655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,1536,0.006735111276308696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,51200,0.8004604445563422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,1024,0.006450666735569636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,512,0.010265777508417765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,12288,0.08000177807278104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,6144,0.012364444633324942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,2048,0.009425777528021071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,2560,0.008168000313970778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,65536,0.9352373547024198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,3072,1.0738062328762479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,16384,0.02811466654141744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,3584,0.03474400109714932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,16384,0.013583111266295115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,4096,0.04001422060860528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,65536,1.166086196899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,128,0.010515555739402771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,4096,0.014311111635631986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,512,0.0038693332009845306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,32,0.0029004443850782183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,2560,0.09203644593556722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,1024,0.1331511073642307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,8192,0.012989333106411828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,12288,0.01567555632856157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,256,0.005108444641033809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,256,0.006568888823191325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,65536,0.5628986888461643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,64,0.030757334497239854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,2560,0.010256888965765635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,3072,0.029290666182835896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,32,0.01955111159218682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,12288,0.02239911092652215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,7168,0.013372444444232516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,256,0.0063591111037466265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,5120,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,1536,0.00813066628244188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,1024,0.021960000197092693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,12288,0.06236088938183255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,65536,0.08384533060921563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,7168,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,3072,0.023577777875794306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,16384,0.6216008928087022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,10240,0.08172444502512614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,65536,1.4613022274441188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,7168,0.03724266754256354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,16384,0.2301164468129476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,768,0.0041964443193541635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,3072,0.013530666629473368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,51200,1.7672435972425673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,6144,0.707497755686442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,2048,0.018058665924602084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,3584,0.009146666361225976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,1536,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,512,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,768,0.017418666018380057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,6144,0.009834667046864828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,128,0.006039111150635614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,51200,0.17931911680433485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,12288,0.13527466191185847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,3584,0.09106666511959499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,1024,0.03587911195225186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,3072,0.01797777745458815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,16384,0.022533333963818018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,6144,0.20498311519622803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,256,0.006436444405052397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,2560,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,3072,0.014304889572991265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,7168,0.01292266696691513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,3072,0.08234755860434638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,256,0.00444088876247406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,32,0.005055111315515307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,128,0.0034648889882696998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,65536,0.012812444733248817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,5120,0.030777778890397813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,16384,0.0940284464094374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,32,0.004595555365085602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,5120,0.07758044534259372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,3072,0.029038220643997192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,1536,0.014112000664075216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,2560,0.008026666939258575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,2048,0.008319999608728621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,2560,0.008930666579140557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,5120,0.009436444275909001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,10240,0.1459653377532959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,7168,0.03321866525544061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,1024,0.009298666483826106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,12288,0.009162666896979014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,8192,0.05183555682500204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,65536,3.6753760443793406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,256,0.004423111263248655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,12288,1.097641732957628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,8192,0.03984977801640829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,1024,0.016421332955360413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,12288,0.042817778057522245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,5120,0.10173155864079793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,51200,0.22767912016974556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,5120,0.08297244707743327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,2560,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,768,0.062260442309909396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,51200,5.101961771647136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,51200,0.05180977781613668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,8192,0.15670843919118246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,768,0.014271999398867289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,5120,0.01149511088927587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,12288,0.9965413411458334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,768,0.02717422280046675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,3072,0.0946542223294576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,8192,0.012171555724408893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,1536,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,10240,0.019409777389632333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,7168,0.026752000053723652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,256,0.0041662222809261745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,8192,0.06102310948901706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,2560,0.006151111175616582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,768,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,3584,0.2650390995873345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,7168,0.02030311193731096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,5120,0.01236977759334776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,5120,0.008192000289758047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,12288,0.4712470902336969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,10240,0.09165777762730916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,2048,0.028024001253975764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,256,0.005274666680230035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,4096,0.00834222220712238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,1536,0.009175111022260454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,2048,0.006591111007663939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,65536,0.09488355451160008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,3072,0.02163911031352149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,6144,0.02290222214327918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,128,0.009574222067991892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,3072,0.6982888645595975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,1024,0.08171288834677802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,5120,0.22065956062740752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,512,0.009926222264766693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,10240,0.017384888397322763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,6144,0.4523937967088487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,2048,0.0151591118839052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,3072,0.0816391110420227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,65536,4.750767178005643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,2560,0.00943288869327969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,65536,0.2091591093275282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,65536,0.0750817788971795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,7168,0.09380444553163315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,2048,0.027764444549878437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,1024,0.4913599756028917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,3072,0.01351377781894472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,51200,4.389478047688802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,16384,0.0510444442431132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,2048,0.006138666636413998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,7168,0.1260319948196411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,768,0.0057795556883017225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,5120,0.04767022199100918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,32,0.0032391111469931076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,3072,0.03983111182848612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,64,0.013814222481515674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,3584,0.03388533327314589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,256,0.017444445027245414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,64,0.16237244341108534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,51200,0.27054399914211696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,3584,0.010791111323568555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,16384,0.03658488723966811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,51200,0.02436977790461646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,4096,0.21292444070180258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,512,0.012824888858530255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,7168,0.186335113313463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,65536,3.6496730380588107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,3072,0.008258666429254744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,2560,1.7213768429226344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,128,0.010440000229411654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,2048,0.021664000219768945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,5120,0.011365333365069495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,3072,0.03298222356372409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,4096,0.04159199860360887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,51200,0.07431288560231526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,64,0.017305778132544625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,128,0.005922666854328579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,6144,0.08900533119837443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,5120,0.04884711239073011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,65536,0.1164284414715237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,16384,0.5091546906365289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,4096,0.01700177788734436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,3072,0.008773333496517604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,2048,0.0188044442070855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,3072,0.011902222202883827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,5120,0.04055644406212701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,8192,0.036383999718560114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,12288,0.017337777548366122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,12288,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,1024,0.0052168890833854675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,768,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,32,0.005365333209435145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,2048,0.00832711077398724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,1024,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,768,0.013198222551080914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,6144,0.04827377862400479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,6144,0.02568622264597151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,6144,0.012879110872745514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,2048,0.00703022215101454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,3072,0.09303110837936401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,7168,0.014308444327778287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,4096,0.006672888994216919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,10240,0.15039288997650146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,512,0.005583999885453119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,10240,0.01184800018866857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,512,0.01833066675398085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,1536,0.011355555719799466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,16384,0.03413955701722039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,3584,0.011423110961914062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,768,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,2048,0.005247999810510212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,3072,0.007233777807818518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,8192,0.30535732375250924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,256,0.01576622161600325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,512,0.0032808888289663526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,65536,0.06952533457014295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,4096,0.026127111580636766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,128,0.004409777621428172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,2048,0.06332177586025663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,1024,0.045203556617101036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,32,0.011418666276666852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,16384,0.2683102289835612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,7168,0.052816000249650746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,1024,0.24776710404290092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,2048,0.006659555352396435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,1024,0.00480266660451889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,1536,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,2048,0.008009778128729926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,256,0.0057546666098965546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,16384,0.023919999599456787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,65536,3.308728959825304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,10240,0.017843556072976854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,3072,0.009185777770148383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,3072,0.03193155593342251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,12288,0.39113510979546445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,512,0.009024889104896123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,128,0.0069653333889113525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,1536,0.1828248898188273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,4096,1.832214143541124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,2048,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,7168,0.09071289168463813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,6144,0.1966773271560669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,3584,0.08047911193635729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,2048,0.008004444340864817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,256,0.005004444470008214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,10240,0.030791110462612573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,51200,12.758013407389322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,512,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,64,0.003568888952334722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,2048,0.08285155561235216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,5120,0.007367110914654202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,12288,0.01978400018480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,3584,0.017996443642510306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,16384,0.025144888295067683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,4096,0.011724444727102915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,6144,0.031009776724709406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,768,0.004184000194072723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,4096,0.015206222732861837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,12288,0.050379554430643715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,256,0.00313688897424274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,64,0.0032666667054096856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,12288,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,3584,0.009546666509575313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,1536,0.019720888800091214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,3072,0.009130666653315226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,16384,0.658640914493137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,8192,0.013340444200568728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,10240,0.7604204283820258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,3584,2.9442115359836154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,5120,0.35946223470899796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,128,0.010065777434243096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,10240,0.013429332938459186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,1536,0.008608889248636033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,4096,0.01237511138121287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,64,0.005409777578380372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,256,0.008404444489214156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,16384,0.02071555621094174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,128,0.010823110739390055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,2048,0.006540444576077991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,256,0.012245333029164208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,7168,0.023992000354660883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,768,0.009248889154858058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,7168,1.8418382008870442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,10240,0.021702221698231165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,4096,0.5980186462402344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,128,0.0340320004357232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,4096,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,6144,0.018073777357737224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,10240,0.023036445180575054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,6144,0.1050800018840366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,64,0.003063111048605707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,32,0.0027751111321979095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,1536,0.013980444934633044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,10240,0.022515555222829182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,5120,0.12270489003923203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,12288,0.04570933183034261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,65536,0.4956746631198459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,65536,3.1585733625623913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,6144,0.03902933332655165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,768,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,1024,0.006213333457708359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,2560,0.050402667787339955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,4096,0.008231999973456064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,3072,0.010830221904648675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,256,0.0035546666218174826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,4096,0.013766222529941134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,6144,0.08791910939746433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,256,0.032074666685528226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,7168,0.0418853329287635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,1536,0.02264533274703556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,16384,0.5303742090861002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,512,0.008343111309740279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,128,0.0070879997478591064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,256,0.00444088876247406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,768,0.00961511085430781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,8192,0.023265777362717524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,8192,0.015088000231319003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,4096,0.01181777815024058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,10240,0.011747555600272285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,2560,0.05106400118933784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,128,0.014561777313550314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,2560,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,768,0.03360177742110358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,64,0.008788444101810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,2048,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,7168,5.128062354193793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,128,0.008003555238246918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,8192,0.14188267125023737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,6144,0.02014666630162133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,1024,0.02574133376280467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,2048,0.01038577738735411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,10240,0.04059022333886888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,16384,0.09105066458384196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,65536,0.042506668302747935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,1024,0.006241777704821692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,8192,0.020037333170572918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,16384,4.4287189907497835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,768,0.08062489165200128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,64,0.13085156016879612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,7168,0.10241599877675374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,256,0.008461332983440822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,2048,0.11410489347245957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,256,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,16384,0.2279440032111274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,256,0.0041555555330382455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,7168,0.6815306875440809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,12288,0.16988533073001436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,3584,0.0173662222094006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,10240,0.031776888502968684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,512,0.009724444813198512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,256,0.005599110904667113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,6144,0.8136320114135742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,7168,0.044581333796183266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,65536,1.7924052344428167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,3584,0.007236444287829929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,6144,0.03804088963402642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,1536,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,3584,0.008911111288600499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,8192,0.21113510926564535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,1024,0.48053868611653644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,4096,0.00886933339966668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,32,0.006033777776691649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,32,0.0038080000215106537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,16384,0.04445599847369724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,65536,0.21580444441901314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,2560,0.011139555937714048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,51200,0.5124506420559353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,8192,0.015030221806632148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,128,0.0028391112056043413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,65536,0.25121866332160103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,1536,0.08259289132224189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,2048,0.014112888111008538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,51200,42.24578857421875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,10240,2.606602562798394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,768,0.007551111280918121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,16384,0.023974221613672044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,5120,0.12958755758073595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,5120,0.022513777017593384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,512,0.008603555460770925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,512,0.005103111267089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,12288,0.017454221844673157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,32,0.0024684444069862366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,5120,0.5665182007683648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,3072,0.024604444702466328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,1024,0.04008888867166307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,32,0.005704000178310607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,512,0.16598933272891575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,3072,0.011132444772455426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,768,0.02032177812523312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,12288,0.01775911119249132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,51200,0.7182675467597114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,6144,0.1636728843053182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,8192,0.1902551121181912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,16384,0.03894311189651489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,3072,0.015024888846609326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,16384,0.2965199947357178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,7168,0.02332088847955068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,3584,0.008588444855478074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,4096,0.021393777595625982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,2048,0.016601777738995023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,768,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,6144,0.27537332640753853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,65536,0.07226400242911445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,2048,0.013412444127930535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,65536,0.17806133959028456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,512,0.05403111047214932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,3584,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,12288,1.6180773841010199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,512,0.02311733365058899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,3584,0.008053333395057255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,10240,0.016001777516471017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,7168,0.02075111038155026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,51200,0.15149777465396458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,32,0.09916266467836167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,51200,0.05259733398755392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,1536,0.006770666688680649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,65536,23.475780910915798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,512,0.031133333841959637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,65536,4.593160841200087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,1536,0.015827554795477126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,7168,0.03399111164940728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,8192,0.03424711028734843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,64,0.010168888502650792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,12288,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,4096,0.07846399810579088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,32,0.002765333279967308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,7168,0.03279377685652839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,12288,0.03057510985268487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,5120,0.008536888493431939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,51200,0.3887253337436252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,10240,0.6722764439053006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,7168,0.14096889230940077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,256,0.0033457777980301115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,6144,0.5453306833902994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,3072,0.00794311116139094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,256,0.011939555406570435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,8192,0.1702373292711046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,768,0.006468444648716185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,6144,0.0702364444732666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,6144,0.6344044473436143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,3072,0.006887110984987683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,3072,0.019651555352740817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,32,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,51200,2.4075039757622614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,8192,0.16669155491722953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,65536,0.3470071156819661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,3072,0.008967111508051554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,5120,0.018399111098713346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,32,0.005570666657553778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,6144,0.010253333383136326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,7168,0.055121776130464345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,51200,1.3448621961805556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,1536,0.007055111229419708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,1024,0.03392444385422601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,768,0.012179555164443122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,65536,10.125601874457464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,1536,0.027760000692473516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,6144,0.8930791219075521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,768,0.01939911146958669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,1536,0.011064888702498542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,16384,0.03226666649182638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,768,0.018046221799320646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,64,0.0028355556229750314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,128,0.014465777410401238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,2560,0.01962311069170634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,16384,0.3032755586836073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,128,0.376127110587226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,128,0.00499199993080563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,5120,0.014825777875052558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,10240,0.039248890346950956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,8192,0.31477067205641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,3072,0.014025777578353882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,3072,0.06379199690288968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,10240,0.048002666897243924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,1536,0.09293955564498901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,12288,0.3619893391927083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,8192,0.012390221986505719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,2560,0.03698044353061252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,32,0.003020444470975134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,65536,0.09430222378836738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,768,0.03150399857097202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,768,0.015843555331230164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,16384,0.026485333840052288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,3584,0.6810275713602701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,768,0.05146400133768717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,256,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,5120,0.14264178276062012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,64,0.002577777744995223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,1536,0.004780444420046276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,7168,0.041745778587129384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,3072,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,16384,0.1479653384950426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,1536,0.009483555952707926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,12288,0.07027822070651583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,256,0.06669688887066312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,2560,0.008813333180215623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,32,0.0025840000145965153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,3072,2.63810306125217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,12288,0.09306844737794663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,4096,0.008438222110271454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,2560,0.03809066613515218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,10240,0.07964089181688097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,5120,0.09703733523686726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,4096,0.006527110934257507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,1536,0.013845333622561561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,128,0.0029493332323100832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,512,0.019127110640207928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,3584,0.0063768890168931745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,256,0.008380444513426887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,2560,0.03696444299485948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,2560,0.011904000408119626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,4096,0.008896888958083259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,6144,0.5660293367173936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,7168,0.1884160041809082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,1024,0.02347822156217363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,4096,0.03726222117741903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,12288,0.006650666809744305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,65536,12.195000542534721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,10240,0.05700977643330892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,64,0.013544888959990608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,65536,0.24815021620856392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,2048,0.009933333430025313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,4096,0.033246222469541765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,6144,0.2102968825234307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,1024,0.04627289043532478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,1024,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,3584,0.07259377506044176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,6144,0.2626390986972385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,51200,0.05730577972200182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,12288,0.06676177846060859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,3072,0.1632044447792901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,10240,0.562252468532986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,12288,0.23827200465732148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,64,0.2270666758219401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,768,0.18971999486287436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,16384,0.061443554030524365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,6144,0.0736515588230557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,12288,0.060623109340667725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,256,0.005205333232879639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,256,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,512,0.003882666842805015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,10240,0.015154666370815702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,128,0.0032044444233179092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,8192,0.020956445071432326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,10240,0.01682222220632765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,1536,0.019926221834288705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,1536,0.009355555805895064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,3584,0.011086222198274402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,4096,0.06558755371305677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,1536,0.0254711111386617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,6144,0.05816977553897434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,128,0.002936888900068071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,2048,0.10284888744354248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,8192,0.023256889647907678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,768,0.0036640001667870414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,10240,0.01611555616060893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,8192,0.009739555418491364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,12288,0.0905253357357449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,7168,0.01812177730931176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,10240,1.1171848509046767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,2560,0.010852444503042432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,256,0.0034542222403817703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,64,0.13160444630516901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,2560,0.05188888973659939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,10240,0.015513777732849121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,51200,0.0496693319744534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,1024,0.12484888235727946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,256,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,3072,0.0075431110130416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,1024,0.0185199992524253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,512,0.05120266808403862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,512,0.009296889106432596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,768,0.019301333361201815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,3072,0.009407999614874521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,51200,0.0863200028737386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,64,0.0052800000541739995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,32,0.00498488876554701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,2048,0.19175022178226045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,512,0.005730666634109285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,6144,0.03104089034928216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,65536,0.181295116742452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,512,0.033583111233181424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,256,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,128,0.003133333391613431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,5120,2.500843471950955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,65536,0.0992444422509935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,128,0.007310222420427535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,2048,0.03440533412827386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,512,0.016081778539551627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,6144,0.02016799979739719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,256,0.0058000000814596815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,2048,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,6144,0.19961244530147979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,3072,0.010855110983053843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,1536,0.05015733175807529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,256,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,32,0.00508977762526936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,16384,0.026523555318514507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,10240,0.006758222149478064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,768,0.007320000065697565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,256,0.0034968890249729156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,1024,0.5812497668796116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,768,0.048467556635538735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,512,0.007724444071451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,32,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,12288,0.02416533397303687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,1536,0.03692888882425096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,6144,0.024640000528759424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,3584,0.02032533288002014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,7168,0.019695111446910434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,5120,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,3584,0.20299200216929117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,768,0.003549333247873518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,768,0.015602666470739575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,3584,0.006681777950790193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,51200,0.6994906531439887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,128,0.0041831110914548235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,2560,0.012879999975363413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,65536,0.11742133564419216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,3072,0.5117271211412218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,768,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,64,0.002732444347606765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,32,0.019133332702848647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,4096,0.36124089029100204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,1536,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,65536,3.69883304172092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,1536,0.005778666585683823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,6144,0.0880791081322564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,12288,0.04489244355095757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,2048,0.07648888561460707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,10240,0.017324444320466783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,3584,0.020418667131000094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,768,0.013738666971524557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,3584,0.011485333244005838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,3584,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,768,0.01651022169325087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,256,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,7168,0.11994043986002605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,3584,0.3376595444149441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,1024,0.010096000300513374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,128,0.008153777983453538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,3584,0.07807822359932794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,51200,0.22770934634738496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,65536,1.778598149617513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,128,0.01256533298227522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,128,0.05731911129421658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,128,0.009697777529557547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,5120,0.4919857978820801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,1536,0.029564443561765883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,8192,0.052074667480256825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,12288,0.14234311050838894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,1024,0.004579555657174853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,12288,1.3015564812554252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,128,0.010016000105275048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,768,0.035199112362331815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,10240,0.08450666401121353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,8192,0.7526328828599719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,12288,0.35967644055684406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,32,0.0026231110095977783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,5120,0.09030399719874065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,3584,0.08242577976650663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,7168,0.014101333088344999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,16384,0.029995554023318823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,51200,20.918782552083332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,3584,0.18140444490644667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,8192,0.4388711187574599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,512,0.005716444303592046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,256,0.012231999801264869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,16384,0.09086666504542033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,5120,0.06690400176578097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,3072,0.03775022096104092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,8192,0.030231111579471167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,51200,0.055014222860336304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,512,0.0034311111602518293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,128,0.05515644285413954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,768,0.010141332944234213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,256,0.004325333154863781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,2560,0.12004800637563069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,32,0.023824888798925612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,65536,0.1445608933766683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,768,0.007999999655617608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,2048,0.012319111161761813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,1536,0.24565333790249297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,51200,8.128025478786892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,4096,0.015896000795894198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,51200,0.7732968860202365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,16384,0.29425332281324595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,16384,0.01921866668595208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,2560,0.011388444238238864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,512,0.0035128887328836652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,1536,0.023639111055268183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,8192,0.04211466511090597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,4096,0.006243555496136348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,8192,0.011810666984981961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,32,0.005161777966552311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,3072,0.017533333765135873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,1024,0.031818668047587075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,1024,0.010315555665228102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,65536,0.8325120078192817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,1024,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,128,0.0041857775714662345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,4096,0.07665244738260905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,16384,0.028015110227796767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,1024,0.01608444419172075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,12288,2.992279900444878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,128,0.004466666529575984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,51200,0.04722222354676989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,10240,0.15768088234795466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,512,0.006667555620272954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,12288,0.015801777442296345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,5120,0.1576391061147054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,2048,0.010403555300500657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,128,0.004436444491147995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,6144,0.40886934598286945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,10240,0.07367910941441853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,1024,0.018391999933454726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,128,0.004570666700601578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,51200,0.19187022580040825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,7168,0.056701335642072886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,12288,0.05422311027844747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,1536,0.012870222330093384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,7168,0.011712888876597086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,1536,0.04204888807402717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,5120,0.026181333594852026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,4096,0.5845244195726182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,51200,14.862491183810763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,12288,0.1606773270501031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,3584,0.03816533419820998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,4096,0.027093332674768236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,1024,0.0058222222659322954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,2560,0.009956444303194681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,512,0.0031946665710873077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,8192,1.0243404176500108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,65536,0.08241866694556342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,3584,0.009943111075295342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,3584,0.010728889041476779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,7168,3.5607155693901906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,64,0.008867555194430882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,768,0.004834666848182678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,10240,0.21393599775102404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,1536,0.0919297801123725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,1536,0.006591111007663939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,3072,0.03562044435077243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,16384,0.021983999345037673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,256,0.011922666596041786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,7168,0.017130666308932833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,768,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,51200,189.67062717013889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,256,0.003649777836269803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,256,0.0033555556502607134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,128,0.01699822147687276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,1536,0.02159022291501363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,12288,0.08235555224948458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,1536,0.010635555618339114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,16384,0.058020446035597056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,1024,0.005076444397370021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,12288,0.011995555626021491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,16384,0.7512657907274035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,10240,0.9962186813354492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,256,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,128,0.0069573331210348345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,768,0.006928000185224745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,6144,0.006777777853939269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,10240,0.641141308678521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,65536,0.07175911135143705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,6144,0.11169511742062038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,3584,0.03163999981350369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,8192,0.08040355311499701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,1536,0.049280888504452176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,4096,0.035712000396516584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,8192,0.02195466723706987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,6144,0.11713688903384739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,10240,0.026296888788541157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,1024,0.012005333271291522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,10240,0.25030843416849774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,6144,1.3910515043470595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,7168,2.9571974012586804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,10240,2.437371571858724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,7168,0.06664355595906575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,5120,0.017424000634087455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,3072,0.015372445185979208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,256,0.8011502159966363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,512,0.003436444534195794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,5120,0.3348551061418321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,2048,0.013337777720557319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,1536,0.06281422244177924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,7168,0.22835644086201987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,4096,0.01797422269980113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,65536,0.0743502246008979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,1536,0.03976088762283325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,8192,0.21527822812398276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,768,0.004753777964247597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,3072,0.0508266654279497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,16384,0.06254488892025419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,12288,0.031501332918802895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,10240,0.01368622233470281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,12288,1.0792640050252278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,3072,0.08321244186825223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,5120,0.11141955852508545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,2048,0.07182666990492079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,768,0.0058319999112023254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,1536,0.020473778247833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,10240,0.0890986654493544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,2560,0.006726222319735422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,3584,0.06673244635264079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,6144,0.11566932996114095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,32,0.005118222286303838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,3072,0.03397599856058756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,768,0.0036799998746977914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,64,0.005191111316283544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,5120,3.8898904588487415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,768,0.022897778285874262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,128,0.015119999647140503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,12288,0.01963911122745938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,1536,0.004671111288997862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,2048,0.007251555720965068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,512,0.0047973332305749255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,512,0.005707555760939916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,2048,0.05503733290566338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,3072,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,65536,0.2634826766120063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,3072,0.2853368918100993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,2560,0.015259555644459195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,6144,21.99910820855035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,64,0.0032960000551409195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,51200,0.47039376364813906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,6144,0.021291555629836187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,128,0.027464000715149775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,128,0.015680889288584392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,8192,0.01145066652033064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,7168,0.014198222094111972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,12288,0.27137600051032174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,12288,0.01819022165404426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,6144,0.01641688909795549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,256,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,256,0.0052915554907586836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,256,0.0036924444139003754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,32,0.0042160000238153665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,1024,0.012175111307038201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,1024,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,512,0.021727111604478624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,3072,0.008799999952316284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,3584,0.012888000243239932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,1536,0.006114666660626729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,1024,0.08585600058237712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,128,0.0041777777175108595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,256,0.02701688806215922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,2048,0.011767999993430244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,4096,0.19785688983069527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,10240,0.055946667989095054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,8192,0.013391111459996967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,6144,0.061456885602739125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,768,0.060497780640920006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,256,0.011535999675591787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,7168,0.013790222505728403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,256,0.012836444709036084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,64,0.0064560001095136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,2048,0.023198222120602924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,12288,0.05275111065970527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,2560,0.322979556189643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,4096,0.014180445008807711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,3584,0.48805777231852215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,16384,0.7840524249606662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,10240,0.4601128896077474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,64,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,7168,0.019130667050679524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,7168,0.007027555671003129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,128,0.0033475555893447665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,256,0.005976888868543837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,2048,0.020502222908867728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,1024,0.004184000194072723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,512,0.0032666667054096856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,2048,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,51200,0.19140444861518013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,6144,0.013840888937314352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,3072,0.04989955491489834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,32,0.005049777941571342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,64,0.045815110206604004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,1024,0.010782221953074137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,12288,0.3975831137763129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,12288,0.03682933251063029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,64,0.005168000029193031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,10240,0.03803911142879062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,4096,0.007857778006129796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,2048,0.015588444140222339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,6144,0.024618667032983568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,8192,0.07449866665734185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,6144,0.015751111838552687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,1024,0.016526222229003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,768,0.004219555606444676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,3584,0.011597333682907952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,8192,0.11701777246263291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,10240,0.01352266636159685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,51200,0.07449066638946533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,256,0.008211555580298105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,128,0.003128888913326793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,1024,0.030123554997973975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,8192,0.04003466831313239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,1536,0.051910221576690674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,6144,0.034859554635153875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,1536,0.46068535910712344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,12288,0.021037333541446265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,6144,0.1229413350423177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,12288,0.39083022541469997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,6144,0.03633066680696275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,3072,0.05957511398527357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,65536,0.25132976637946236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,1024,0.0040755557517210645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,3072,0.007340444458855524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,6144,0.5113591088189019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,65536,20.221822102864582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,3584,0.01447111037042406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,51200,0.01442311041884952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,51200,0.23359288109673393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,2048,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,6144,0.0425911115275489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,3072,0.007542221910423703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,16384,0.4889101982116699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,7168,0.017629333668284945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,65536,0.03974311219321357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,2048,0.011516444385051727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,7168,0.012714666624863943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,6144,0.3397502104441325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,1536,0.01531644496652815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,51200,0.0319164428446028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,51200,0.04345600141419304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,7168,0.021107556091414556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,6144,0.0123511114054256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,3584,0.007802666889296637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,512,0.0658444431093004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,2048,0.0053377776510185665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,1024,0.02125955621401469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,3072,0.008590222232871586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,1536,0.21909688578711617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,64,1.0625680287679036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,8192,0.2924293412102593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,8192,0.06495822138256498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,5120,0.026313778426912095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,65536,0.9175412919786242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,128,0.015222221612930298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,3072,0.012303111453851065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,7168,0.10389600197474162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,65536,1.734826617770725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,7168,0.15726844469706217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,64,0.0031022222505675424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,12288,0.04360533422893948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,5120,0.007191110816266801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,768,0.006797333144479328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,7168,0.0711502234141032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,3072,0.01571999986966451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,4096,2.5851732889811196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,768,0.009008000294367472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,12288,0.09104533327950372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,7168,0.05464977688259549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,3072,0.11412089400821263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,51200,0.4218293295966254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,16384,0.02351911034848955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,512,0.013616888887352414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,128,0.006102222121424145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,7168,0.058416002326541476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,1536,0.12595200538635254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,2048,0.005433777968088786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,65536,0.1276008950339423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,2048,0.03713333275583055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,768,0.005577777822812398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,12288,0.09760799672868516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,2048,0.22966933250427246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,64,0.00556355549229516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,1024,0.29574399524264866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,10240,0.006256888724035687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,3072,0.008347555167145198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,2560,0.00835111074977451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,3584,0.01864888932969835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,3584,0.03551377852757772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,10240,0.6076542006598579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,2048,0.00980355590581894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,4096,0.275036440955268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,2560,0.02425688836309645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,3584,0.09736533297432794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,65536,0.19262311193678114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,3072,0.016732444365819294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,3584,1.5384462144639757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,6144,0.1864222288131714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,7168,0.4565439754062229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,51200,0.13805600007375082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,1536,0.013390222357379066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,16384,0.7124826643202039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,8192,0.42688711484273273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,2560,0.16900622844696045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,32,0.003699555579158995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,12288,5.2896169026692705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,1536,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,2560,0.009405333134863112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,512,0.1989253362019857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,3584,0.007367110914654202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,16384,0.02067377832200792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,16384,0.04064711266093784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,2048,0.02643644478585985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,64,0.0444524453745948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,256,0.004439110971159405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,256,0.028597331709331934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,5120,0.0221413340833452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,4096,0.04335644510057238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,65536,0.07405244641833834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,6144,0.010932444698280759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,64,0.005102222164471944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,10240,0.04040711124738058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,128,0.0034648889882696998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,3072,0.015056888262430826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,128,0.005495999836259418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,6144,0.3195377720726861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,2048,0.017145777742067974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,1536,0.03385688861211141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,6144,0.10982666412989299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,64,0.00871377769443724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,6144,0.08261244164572822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,32,0.002704888789190186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,65536,0.7340835465325249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,2048,0.007901333272457123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,32,0.023697778582572937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,5120,0.014450665977266101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,1024,0.005886222339338726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,4096,1.4620613521999781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,2560,0.009263111485375298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,65536,0.8720080057779948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,3072,0.01219377749496036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,12288,0.07551822397443983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,10240,0.03925688730345832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,10240,0.017754667335086398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,65536,4.844896104600695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,128,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,7168,0.014525334040323893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,6144,0.03138577938079834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,12288,0.2110053300857544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,12288,0.5328639878167046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,512,0.012271111210187277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,2048,0.015119999647140503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,6144,0.013833777772055732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,12288,0.04267999860975477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,128,0.012721777790122561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,4096,0.006167111297448476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,2048,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,16384,1.016824934217665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,10240,0.015178667174445258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,128,0.013912000589900546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,512,0.021612443857722815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,64,0.011108444796668159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,65536,0.21237066056993273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,8192,0.010899555351999072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,5120,0.2268737687004937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,1536,0.02463200026088291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,3072,0.07117511166466607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,256,0.007830222447713217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,2048,0.04827733172310723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,1536,0.007921777665615082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,512,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,128,0.007125332951545715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,10240,0.006980444822046492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,10240,0.03362222181426154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,3072,0.01902933253182305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,16384,0.04242755638228523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,5120,0.1991999944051107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,128,0.019339554839664035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,128,0.02197155521975623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,768,0.03559289044804043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,12288,0.1956275569068061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,768,0.014548444085650973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,2560,0.027800000376171533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,512,0.004813333352406819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,1536,0.057182219293382436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,5120,0.024382222029897902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,65536,0.1198764509624905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,32,0.005947555518812603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,2048,0.016740444633695815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,3584,0.019075555933846366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,768,0.027758222487237718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,12288,0.29444800482855904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,5120,1.0277609295315213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,3072,0.09933688905504014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,256,0.030300445026821558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,512,0.17618221706814238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,1536,0.0052737775776121355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,10240,0.11356355084313287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,4096,0.03106844425201416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,51200,0.4140284326341417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,16384,0.046535998582839966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,1024,0.010231110784742568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,7168,0.012266666524940066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,51200,0.4577200147840712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,8192,0.03146222233772278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,1536,0.007625777688291337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,32,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,4096,0.04313244422276815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,5120,0.036927110619015165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,8192,0.032398223876953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,1536,0.015769778026474845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,10240,0.01353422221210268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,51200,0.018979556030697294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,65536,0.1403635607825385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,65536,2.905252456665039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,16384,0.03660000032848782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,8192,0.3829457759857178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,1536,0.011928000383906893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,10240,0.26932088534037274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,64,0.02676800058947669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,64,0.019807111885812547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,10240,0.27913154496086967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,6144,0.021031111478805542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,3584,0.015081778168678284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,16384,0.11843821737501357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,51200,0.40919110510084367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,4096,0.05952533086140951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,64,0.0028026666906144884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,12288,0.0192586663696501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,5120,0.006843555718660355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,4096,0.026027555267016094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,2048,0.0203146669599745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,51200,0.12533333566453722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,1024,0.01572444372706943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,12288,0.10645866394042969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,512,0.005024888863166173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,32,0.017041777571042378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,64,0.003310222178697586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,6144,0.019095111224386428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,6144,0.21089066399468315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,51200,16.315655178493923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,7168,0.016388444436921012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,2560,0.029735998974906072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,512,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,1536,0.00674222202764617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,768,0.004219555606444676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,7168,0.02528000043498145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,128,0.02719466719362471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,2048,0.0504755543337928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,2560,0.4553786913553874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,8192,0.027765333652496338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,7168,0.028100444210900202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,6144,0.0732826656765408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,2048,0.02010222276051839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,2560,0.0499511096212599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,512,0.010694222317801582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,64,0.003947555604908201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,5120,1.2485991583930123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,2048,0.09041333198547363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,16384,0.010501333408885531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,4096,0.011585777832402123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,256,0.011866666376590729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,3072,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,8192,0.1423528856701321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,768,0.02773155603143904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,64,0.002514666567246119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,7168,0.7103786468505859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,2048,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,6144,0.0101742222905159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,10240,0.015496888094478183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,10240,0.025553777813911438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,65536,10.26391855875651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,3072,0.0447377794318729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,10240,0.022841778066423204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,2560,0.09333155552546184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,51200,0.025946666797002155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,256,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,2048,0.012318222059143914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,1536,0.07728888591130574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,1536,0.006025777922736273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,256,0.013367999758985309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,2560,0.11400889025794135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,2048,0.07878222068150838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,5120,0.014249778456158109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,8192,0.015071110592948066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,2560,0.028852442900339764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,3584,0.11771021948920356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,1024,0.003922666526503033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,1536,0.00479288895924886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,2048,0.19526399506462944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,5120,0.009507555928495195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,3072,0.19489599598778617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,5120,0.12184711297353108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,16384,0.026200888885392085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,512,0.0032391111469931076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,51200,0.011807999677128263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,5120,0.08090666929880778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,7168,0.0447253319952223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,256,0.0027306667632526825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,768,0.005592888842026393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,10240,0.6555297639634874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,5120,0.03190755512979295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,2048,0.7076240115695529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,3584,0.12083555592430963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,8192,0.006009777800904379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,4096,0.011485333244005838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,12288,0.07937333318922254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,3072,0.011657777759763928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,2560,0.14571821689605713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,512,0.017059556312031217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,12288,0.029857777886920508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,5120,0.2191751135720147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,51200,0.08452800247404312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,768,0.006165333506133821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,3584,0.008183110919263627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,8192,0.026552889082166884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,1024,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,128,0.06527822547488742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,3584,0.04844355583190918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,32,0.0052728888889153796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,3584,0.0993751088778178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,2048,0.018914666440751817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,1024,0.018440888987647164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,65536,2.444176991780599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,768,0.03037066592110528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,65536,0.12603110737270778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,2560,0.007737778127193451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,1024,0.03181511163711548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,1536,0.016318221886952717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,3584,2.190464867485894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,256,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,256,0.005218666460778978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,128,0.015236443943447538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,6144,0.02568888829814063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,512,0.006413333531883027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,6144,0.027791112661361694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,5120,0.10709244675106472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,64,0.002611555573013094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,1024,0.00814577771557702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,10240,0.07352533605363634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,8192,0.19836889372931588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,3072,0.02962755494647556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,3072,0.014703111516104804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,12288,0.017048888736300997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,512,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,512,0.13102844026353624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,768,0.005803555664088991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,7168,0.01091822236776352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,512,0.013515555196338229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,32,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,2048,0.008486222061845992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,16384,0.086008886496226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,3072,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,32,0.005047111047638787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,1024,0.022220444348123338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,4096,0.01826755536927117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,3072,0.09678755866156684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,3584,0.008440000315507254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,5120,0.026895999908447266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,51200,0.05450310972001818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,1536,0.010967999696731567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,64,0.0026062221990691293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,1536,0.0058124446206622655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,6144,0.0184560004207823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,10240,0.026800000005298193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,256,0.006984000404675801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,2560,0.009586667021115622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,7168,0.01072177787621816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,6144,0.2824844519297282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,768,0.023786667320463393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,2560,0.010816888676749336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,2560,0.010287111004193624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,6144,0.01592266725169288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,16384,0.24186756875779894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,768,0.007365333537260692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,32,0.002570666579736604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,4096,0.04433600107828776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,16384,0.17107821835411918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,768,0.005658666706747479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,8192,0.18911822636922201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,6144,0.04641333222389221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,1536,0.013827555709415011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,51200,0.23208443323771158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,1536,0.011401777466138205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,16384,0.44813330968221027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,12288,3.620283762613932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,7168,0.01367288910680347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,10240,0.177236451043023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,7168,0.021354667014545862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,8192,0.03937600056330363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,12288,0.018423111902342904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,2048,0.011766222616036734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,256,0.007156444092591603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,16384,0.018395556343926322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,2560,0.861951986948649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,8192,0.025825776987605627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,6144,0.041527112325032554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,3584,0.018191999859280057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,1536,0.03619199991226196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,16384,0.5505039956834581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,64,0.005606222069925732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,512,0.005984000033802456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,16384,0.06155200137032402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,2560,0.05841155846913656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,3584,0.042430222034454346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,5120,0.15890399614969888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,128,0.006029333505365584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,16384,0.368943108452691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,6144,0.02416533397303687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,10240,0.018561777141359117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,12288,0.21310577127668592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,1024,0.01053066634469562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,7168,0.014831999937693277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,2048,0.011445333560307821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,512,0.008488888541857401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,51200,0.00963466697269016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,65536,4.901054382324219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,65536,7.356988694932725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,128,0.003952888978852166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,1024,0.007426666716734569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,7168,0.01683555543422699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,256,0.020076443751653034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,512,0.02441155579355028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,10240,0.2753404511345757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,16384,0.12530933486090765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,2048,0.13087556097242567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,1024,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,12288,0.030031111505296495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,1536,0.062223109934065074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,64,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,3584,0.008470222353935242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,128,0.004829333474238713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,3072,0.07332622342639498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,256,0.01238044434123569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,3072,0.04808444446987576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,1024,0.009917333722114563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,2560,0.013730666703648038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,32,0.0029093333416514923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,12288,0.09053955475489299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,4096,0.0839937792883979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,51200,0.04487733377350701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,10240,0.017439999514155917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,3584,0.07326400279998779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,64,0.005321777943107817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,4096,0.2208871046702067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,128,0.004105777790149053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,128,0.015623110863897534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,6144,0.014516444669829475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,512,0.003826666623353958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,64,0.003810666501522064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,2560,0.10167911317613389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,64,0.020346666375796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,512,0.036357333262761436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,3584,0.09135466814041138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,6144,0.026055110825432673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,8192,0.07545777824189928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,512,0.007753777835104201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,51200,0.5620826615227593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,64,0.006711999989218182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,8192,0.013259555730554791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,2048,0.011529777612951068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,10240,0.021635555558734473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,1536,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,51200,2.0171616872151694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,3584,0.009761778016885122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,1024,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,12288,0.6748977767096626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,768,0.01626488897535536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,3072,0.016579555140601266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,10240,8.54037136501736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,2560,0.0755537748336792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,16384,0.07209510935677423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,8192,1.013612429300944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,65536,0.09400088919533624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,256,0.005730666634109285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,8192,0.017154667112562392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,16384,0.01833777791923947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,32,0.10275199678209092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,51200,0.5147626664903429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,512,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,3072,0.009352000223265754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,6144,0.11428533660040961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,16384,0.45934756596883136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,2048,0.008151999778217739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,65536,0.0894426637225681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,1024,0.09866488642162746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,3584,0.008151999778217739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,1536,0.009897777603732215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,2048,0.021931555536058214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,64,0.005225777626037598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,2560,0.12591111660003662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,65536,0.1390719943576389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,4096,0.043077333105934985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,3584,0.01568622224860721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,2560,0.017260443833139207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,3584,0.01696266730626424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,256,0.004252444538805219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,2048,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,2560,0.01069155583779017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,3072,0.07315911187065972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,32,0.019173334042231243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,1024,0.053745776414871216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,3072,0.17630222108629015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,4096,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,32,0.00927466650803884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,768,0.04522844486766391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,32,0.008419555922349295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,1024,0.009956444303194681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,6144,0.04136711027887132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,8192,0.01143022212717268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,5120,0.006280888699822956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,2560,0.005919111271699269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,16384,0.04813244607713488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,256,0.004787555585304896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,32,0.002571555475393931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,768,0.005936000082227919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,128,0.004800888813204235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,512,0.03991733325852288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,768,0.010838222172525195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,3072,0.016685333516862657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,64,0.005741333381997214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,3584,0.023891554938422307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,64,0.04390133420626322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,10240,0.016813332835833233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,16384,0.1609626611073812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,1024,0.009597332941161262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,2560,0.04196444484922621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,4096,10.24088372124566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,16384,0.12559910615285239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,1024,0.013032889200581444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,51200,1.0527849197387695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,10240,0.3345199955834283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,6144,0.04227022329966227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,51200,0.06601155466503568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,2560,0.006176888942718506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,65536,0.1869386699464586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,128,0.009429333110650381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,6144,0.30292799737718373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,1536,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,4096,0.015190222197108798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,1024,0.031039112144046362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,12288,0.017408889200952318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,2048,0.05292266607284546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,6144,0.24260799090067545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,256,0.005555555638339784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,2560,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,2048,0.08705155717002021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,64,0.005159111072619756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,65536,1.9954132503933375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,1024,0.01460355520248413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,768,0.007799111306667328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,51200,2.159607145521376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,1024,0.004807999978462855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,3072,0.009924444059530893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,7168,0.01692800058258904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,1024,0.006223111102978389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,8192,0.03993777765168084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,2560,0.01904977858066559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,51200,0.11280977725982666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,128,0.0029973333908451926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,16384,0.06708355744679768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,12288,0.016511110795868766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,1024,0.006725333217117522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,512,0.01901955571439531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,768,1.5404604805840387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,65536,4.692654927571614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,64,0.0033688888781600525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,4096,0.4928444226582845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,4096,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,8192,0.7414844301011828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,256,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,768,0.0069120000633928515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,4096,0.07107644610934787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,3584,0.009280000295903947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,2048,0.017196445001496207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,3584,0.014484445254007975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,2560,0.5184648831685384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,51200,0.3596151140001085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,3072,0.014583110809326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,512,0.005456888841258154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,32,0.0037128888070583344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,512,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,8192,0.09148266580369736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,2048,0.02149066660139296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,5120,0.030761778354644775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,65536,1.7374773025512695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,2560,0.039437333742777504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,12288,0.2685200108422173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,8192,0.020023999942673575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,128,0.0037573331760035623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,6144,0.013639111485746173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,128,0.0027751111321979095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,4096,0.032003555032942034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,8192,0.2843395604027642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,3584,0.008259555531872643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,1536,0.0064239998658498125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,16384,0.07522666454315186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,4096,0.06679466697904798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,2048,1.3449796040852864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,16384,0.5640453232659234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,1536,0.014008000493049622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,1536,0.007611555357774098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,8192,0.016892444756295946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,7168,0.031194667021433514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,5120,0.010109333528412713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,5120,0.10951822333865696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,8192,0.0437520013915168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,12288,0.20723199844360352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,768,0.004970666848950916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,16384,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,2560,0.020035554965337116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,5120,0.2080897755093045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,512,0.058280891842312284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,64,0.005240889059172736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,1536,0.18287289142608643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,2560,0.008449777960777283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,5120,0.29415912098354763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,65536,5.516452365451389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,7168,0.01755466726091173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,128,0.004925333377387789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,16384,0.05537333422236972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,768,0.02695200012789832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,12288,1.7493261761135523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,6144,0.022462222311231825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,3584,0.011830222275522021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,1536,0.013076444466908773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,16384,0.8981030782063802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,8192,0.016576889488432143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,4096,0.10164711210462783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,6144,0.04938044481807285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,65536,0.17608266406589082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,6144,0.028098666005664404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,1536,0.09804888566335042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,3584,0.016143111719025504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,768,0.009104889300134447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,8192,0.008632000121805403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,4096,0.14098933007982042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,256,0.004599110947714912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,10240,0.05861688984764946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,6144,0.009697777529557547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,4096,0.017882666654056974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,2560,0.007853333320882585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,6144,0.019317333896954853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,51200,6.396760728624131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,512,0.0041662222809261745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,16384,0.24718666076660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,4096,0.18142222033606636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,512,0.011315555208259158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,4096,0.6415048705206977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,3584,0.011608888705571493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,6144,0.012269333004951477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,6144,0.26231998867458767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,2560,0.017180444465743173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,7168,0.016499555773205228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,3584,0.3535919984181722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,5120,0.017827555537223816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,768,0.003552000141806073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,1536,0.0203146669599745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,4096,0.010637332995732626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,51200,0.15716621610853407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,2560,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,256,0.021849777963426378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,7168,0.06671466430028279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,6144,0.013167111410035027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,4096,0.010600888894663917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,256,0.004453333301676644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,12288,1.2695929209391277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,32,0.009129777550697327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,51200,0.060496886571248375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,1024,0.02756622268093957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,6144,0.13277867105272081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,768,0.007429333196745978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,64,0.002756444530354606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,256,0.029248889949586656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,7168,0.07990399996439616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,1024,0.0587057802412245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,10240,0.01423555612564087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,512,0.008282666405042013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,6144,0.24077955881754556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,1536,0.0063697778516345555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,512,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,768,0.02050844497150845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,10240,0.02750844425625271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,128,0.0052693333062860705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,768,0.006243555496136348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,3072,0.009890666438473595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,1024,0.00574933323595259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,10240,0.015957333975368075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,6144,0.07714933156967163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,6144,0.10785155826144749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,768,0.005167110926575131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,64,0.00500266667869356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,32,0.0032133333798911837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,5120,0.12029955122205947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,1536,0.011271110839313932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,2560,0.007123555574152205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,768,0.023175999522209167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,512,0.008991111483838823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,1536,0.012107555237081317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,8192,0.011519999967681037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,8192,0.11201777723100449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,512,0.008594666918118795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,16384,0.06752444638146295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,7168,0.021553777986102637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,4096,0.10690044694476658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,3584,0.025370667378107708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,1536,0.20864266819424102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,2048,0.010355555348926121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,1536,0.01533511115445031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,2560,0.19755289289686415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,65536,0.12845333417256674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,1024,0.008775110873911116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,4096,0.33918576770358616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,512,0.005892444401979446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,12288,3.316449695163303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,16384,1.3226346969604492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,10240,0.20730755064222547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,512,0.0041893331540955436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,64,0.006774222271309958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,51200,0.3181697792477078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,4096,0.02342222299840715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,128,0.04591466652022468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,1536,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,16384,0.05250933435228136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,4096,0.018728888697094388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,256,0.009163555171754625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,7168,0.006194666855865055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,12288,0.01956977778010898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,6144,0.011043555206722684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,128,0.015829333000712924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,2048,0.06381155384911431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,1536,0.005440000030729506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,3072,0.01533333294921451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,512,0.0033555556502607134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,6144,0.029170665476057265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,2048,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,16384,0.015257777439223396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,128,0.005931555396980709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,51200,0.5323848724365234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,32,0.0034515555534097883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,10240,0.33815733591715497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,3072,0.0064035554726918536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,2048,0.032271110349231295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,12288,0.02530844509601593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,10240,0.9407289293077258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,2560,0.13500089115566677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,768,0.006647999915811751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,3584,0.015488889482286243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,7168,0.11654133266872829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,3584,0.01590488851070404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,5120,0.03687199950218201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,512,0.013796444568369122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,16384,0.07840533388985528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,16384,0.5670115682813857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,256,0.021920000513394673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,32,0.003248888792263137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,2560,0.010964444114102257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,4096,0.07971377505196466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,3072,0.14568889141082764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,128,0.0027831111931138565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,4096,0.10478133625454372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,65536,2.6070213317871094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,7168,0.04517777760823568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,1536,0.008192000289758047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,256,0.05859466393788656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,2048,0.11104710896809895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,32,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,32,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,2048,0.007430222299363878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,512,0.19585511419508192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,512,0.0031804444475306403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,512,0.1476328902774387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,8192,0.008374222450786168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,768,0.021914665897687275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,2560,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,6144,0.010341333018408881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,3584,0.011591110792424945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,512,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,256,0.00314666661951277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,1536,0.010419555836253695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,10240,0.0948995550473531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,5120,0.006821333534187741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,5120,0.008822222550710043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,64,0.015825778245925903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,3584,0.060121774673461914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,2048,0.03325600094265408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,3584,0.055953780810038246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,10240,0.04108000132772658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,2048,0.0070168889231152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,64,0.016452444924248587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,8192,0.012171555724408893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,10240,0.027791112661361694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,8192,0.02313333253065745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,65536,0.06608000066545275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,16384,0.06432711415820651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,2048,0.010856889188289642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,3584,0.2518666585286458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,2048,0.05230755607287089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,4096,0.015796444482273526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,7168,0.02818044357829624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,3072,0.13848977618747288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,16384,0.10544088813993667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,1536,0.0046453335218959385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,32,0.026181333594852026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,32,0.005186666631036335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,16384,0.7600648668077258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,512,0.008488000267081791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,4096,0.31361601087782115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,3072,0.011156444748242697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,256,0.004611555486917496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,4096,0.022265776991844177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,7168,0.029799110359615747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,12288,2.507563485039605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,768,0.007416889071464539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,32,0.009400889277458191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,8192,0.04320355587535434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,5120,0.009740444521109263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,4096,0.35218222935994464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,128,0.0028657778683635923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,51200,0.0724124444855584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,2560,0.017978666557206046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,128,0.048467556635538735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,3072,0.009328888522254096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,4096,0.008587555752860175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,512,0.004559111264016894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,3072,0.0074879998962084455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,768,0.004915555732117759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,12288,0.17950310972001818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,256,0.008929777476522658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,65536,0.1057751112514072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,64,0.0026951111439201566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,1536,0.0046924443708525766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,51200,3.6715172661675344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,1024,0.012621333201726278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,51200,0.11936533451080322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,768,0.01093777765830358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,1536,0.1578471130794949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,256,0.003113777687152227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,1536,0.02924444609218174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,3584,0.017296888762050204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,51200,1.1201768451266818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,2048,0.03385333220163981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,5120,0.010139555566840703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,64,0.04551022251447042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,512,0.005034666508436203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,32,0.007134222322040134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,1536,0.050231112374199756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,1024,0.009304000271691216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,64,0.008656889200210571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,51200,0.19697156217363146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,64,0.007335999773608313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,512,0.003941333128346337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,7168,0.08413066466649373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,10240,0.6210604243808323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,16384,7.407331678602431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,768,0.008029333419269985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,128,0.006599999964237213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,16384,0.5753493309020996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,3584,0.008966222405433655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,5120,1.2044844097561307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,2048,0.01572711103492313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,4096,0.011549333731333414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,6144,0.3332657814025879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,12288,0.03357244531313578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,1024,0.01016711112525728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,3584,0.010126222338941362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,16384,0.047777778572506376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,128,0.012276444170210095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,8192,0.021049777666727703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,10240,0.06796444124645658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,128,0.01130488928821352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,1024,0.004388444539573458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,16384,0.7197599940829806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,1024,0.021984888447655573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,65536,4.402951134575738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,3584,0.7729297743903266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,8192,0.641001754336887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,2560,0.05159022079573738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,12288,0.02313866714636485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,2048,0.008198222352398766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,2560,0.0068408888247278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,7168,0.014208000567224292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,5120,1.2902213202582467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,2048,0.0069431112044387394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,32,0.011733333269755045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,5120,0.03706489006678263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,3584,0.015427556302812366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,12288,0.9856986999511719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,128,0.0033075554917256036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,12288,0.8478648927476672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,7168,0.1386622190475464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,3584,0.012744888663291931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,2048,0.044533332188924156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,1024,0.009461333354314169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,16384,0.05253244439760844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,6144,0.020209777686331008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,65536,0.02472711106141408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,2560,0.07610844241248237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,2560,0.02258844508065118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,4096,0.011591110792424945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,4096,0.014880888991885714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,128,0.014105778601434497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,6144,0.11308977339002822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,51200,0.06190133094787598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,3584,0.32311288515726727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,768,0.09004444546169704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,8192,0.4557209014892578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,256,0.007548444800906711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,1024,0.027165333429972332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,3584,0.013978666729397245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,7168,0.020776889390415616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,768,0.01183733344078064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,8192,0.026566222310066223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,256,0.030248890320460003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,5120,0.042972445487976074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,10240,0.5655235714382595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,4096,0.5208720101250542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,512,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,16384,0.034803556071387395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,256,0.0035377778112888336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,1536,0.09249244133631389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,65536,0.3544142246246338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,1536,0.00867377801073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,128,0.07064444488949247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,8192,0.01938488913906945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,5120,1.5578301747639973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,2048,0.010630221830474006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,768,0.005702222386995952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,4096,0.08255733384026422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,2048,0.849182234870063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,1024,0.026544888814290363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,3584,0.01934400035275353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,51200,0.18526221646202934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,256,0.0032799999333090256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,32,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,16384,0.02349600030316247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,16384,0.1284959978527493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,51200,1.6423386467827692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,3072,0.007990222010347579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,7168,0.03046311272515191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,256,0.0036355555057525635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,128,0.0033039999090962936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,8192,0.20661244127485487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,1024,0.009570666485362584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,128,1.8968630896674261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,3584,0.03413510984844632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,1024,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,4096,0.02035555574629042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,12288,0.04056355688307021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,768,0.005025777965784073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,51200,0.4075431029001872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,5120,0.01293333371480306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,2048,0.01979377700222863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,8192,0.022035555707083807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,5120,0.010803555448849997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,256,0.007287999822033777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,10240,0.040596445401509605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,6144,0.12827466593848333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,128,0.004970666848950916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,8192,0.14012622833251953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,3584,0.5640773243374294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,16384,0.5297182401021322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,10240,0.03250311149491204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,51200,0.07123733229107328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,10240,0.12947200404273138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,4096,0.02088533341884613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,512,0.01976088848378923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,12288,0.592648877037896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,1024,0.04943022131919861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,51200,0.10286933183670044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,1024,0.41345779101053876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,6144,2.509535047743056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,7168,0.02457422183619605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,4096,0.019036443697081674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,51200,0.010813333094120026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,2048,0.011894221935007306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,16384,0.1753902170393202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,10240,0.034392890002992414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,512,0.01611022154490153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,8192,0.048375109831492104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,7168,0.01830400029818217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,51200,6.991973453097874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,1024,0.7452533509996203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,256,0.013819555441538492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,64,0.0035822221802340615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,1536,0.00816000004609426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,1024,0.0101742222905159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,3072,0.015016888578732809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,8192,0.020050666398472257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,6144,0.007517333659860823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,256,0.005426666802830166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,4096,0.013895110951529609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,768,0.007576000359323289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,128,0.013794666363133324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,51200,0.05535733368661669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,12288,0.022343110707071092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,16384,0.030142221185896132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,10240,0.044459554884168834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,32,0.0026195556339290407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,10240,0.023963555693626404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,4096,0.009935999910036722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,6144,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,1536,0.07142577568689983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,1536,0.0159608887301551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,10240,0.023600000474188063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,8192,0.016955556141005624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,1536,0.016333333320087857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,10240,0.032087110810809664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,7168,0.076109336482154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,1024,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,1024,0.38084978527492946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,4096,0.13491911358303493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,1536,0.007795555724038019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,7168,0.05295822355482313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,512,0.010515555739402771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,1536,0.013160000244776407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,4096,0.11313511265648736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,768,0.015360888507631091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,1024,0.048362665706210665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,7168,0.05753599935107761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,10240,0.14855733182695177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,51200,0.16684267255995008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,8192,0.18145333396063912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,10240,0.01351199961370892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,65536,7.2541088528103295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,768,0.011690666278203329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,256,0.0036071112586392295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,51200,0.021172444025675457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,5120,0.09892355733447605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,16384,0.39333510398864746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,12288,7.3885693020290795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,6144,0.05727822250790066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,4096,0.019107555349667866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,16384,0.10427288876639472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,2048,0.028997331857681274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,1536,0.03577422102292379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,16384,0.08146399921841092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,6144,2.552320056491428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,8192,0.0361768901348114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,256,0.003814222084151374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,128,0.010987555815113915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,51200,1.7173830668131511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,12288,0.012303999728626676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,5120,0.042133334610197276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,8192,0.10233955250846015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,5120,0.009385777844323052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,64,0.005416000054942236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,12288,0.021328000558747187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,3584,0.009966221948464712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,768,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,4096,0.1359857718149821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,64,0.005670222143332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,2560,0.020303999384244282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,2560,0.01998488936159346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,768,0.04239555531077915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,512,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,512,0.005425777700212266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,3584,0.030696001317765977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,768,0.005799110978841782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,51200,0.2697644498613146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,2048,0.015456888410780164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,51200,0.06067555480533176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,16384,0.06504444281260173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,1536,0.03327466713057624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,64,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,512,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,8192,0.0063946665161185795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,12288,0.03253511256641812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,10240,0.0381297767162323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,2048,0.09133066733678182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,16384,0.12587289015452066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,768,0.007610666255156199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,16384,0.12112889024946426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,16384,0.014045332868893942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,5120,0.021130666136741638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,3584,0.007821333077218797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,16384,0.3245893319447835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,6144,0.026608000199000042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,256,0.12609955999586317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,6144,0.1542817751566569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,8192,0.034352888663609825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,1536,0.006520000182920032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,5120,0.021913778450753953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,16384,0.4760799937778049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,12288,0.1120720042122735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,6144,0.048496001296573214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,1536,0.013677332964208392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,10240,0.24488533867730033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,10240,0.106568886174096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,1536,0.004709333181381226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,512,0.003656889001528422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,51200,0.08688888947168986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,3584,0.014469333820872836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,2560,0.015537778536478678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,8192,0.03062844607565138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,10240,0.01701244380739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,51200,0.23962312274509004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,768,0.007785778078767989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,2560,1.1348870595296223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,12288,0.008779555559158325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,128,0.008905777500735389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,1024,0.054514666398366295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,12288,0.056167112456427686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,6144,0.021764443980322942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,2048,0.009266667068004608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,1536,0.015798222687509324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,1024,0.06993955373764038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,16384,0.021482666333516438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,65536,0.07367466555701362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,8192,0.012351999680201212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,2560,0.021518222159809534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,512,0.0047146665553251905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,51200,0.9997137917412652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,3584,0.0313368903266059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,512,0.0035386664999855887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,2048,0.0075342224703894714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,4096,16.95975579155816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,512,0.0064275554484791225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,6144,0.15907822714911565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,10240,0.023936889237827722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,1024,0.39069332016838926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,512,0.04625422093603346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,2048,0.029912001556820337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,7168,0.025605332520272996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,768,0.010910222099887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,3072,0.00759733302725686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,6144,0.17790488402048746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,16384,0.10878222518497044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,2560,0.15943733851114908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,2560,0.014864888456132678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,6144,0.02798933453030056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,3584,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,2560,0.056271996763017446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,51200,4.550440046522352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,128,0.015191111299726697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,32,0.003109333415826162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,2048,0.20482399728563097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,256,0.0037706668178240457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,32,0.0034737777378824023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,1024,0.00723555518521203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,8192,0.015260444747077094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,51200,0.09356711308161418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,51200,0.035623110002941556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,7168,0.027814222706688776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,10240,0.008912888665994009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,6144,0.12764267126719156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,4096,0.009731555150614845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,4096,0.010800888968838586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,512,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,4096,0.015365334020720588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,12288,0.037456889947255455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,512,0.0162124451663759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,64,0.016851555969980028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,32,0.0027715555495686005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,2560,0.12418577406141494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,12288,0.03930222325854831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,12288,0.928943951924642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,2560,0.027639110883076985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,3584,0.00795911086930169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,32,0.01124533348613315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,1024,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,65536,0.13496711519029406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,2560,0.009189333352777693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,128,0.011468444433477191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,12288,0.015609777635998197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,16384,0.02015466656949785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,1024,0.007333333293596904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,65536,3.0206578572591147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,2048,0.009675555759006077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,65536,0.09827644295162624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,3584,0.015617777903874716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,1536,0.020827554994159274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,7168,0.017961778574519686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,10240,66.12223985460069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,512,0.026123555170165166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,2048,0.009550222092204625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,2560,0.02597155504756504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,16384,0.040004445446862116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,51200,0.0626479983329773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,5120,0.013528000149461957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,7168,0.061526219050089516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,5120,0.031003554662068684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,1024,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,1536,0.09246133433447944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,1024,0.0058106668293476105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,51200,0.10931289196014404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,51200,0.5611288812425401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,2560,0.008960888617568547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,5120,0.013979555832015144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,4096,0.05244444476233589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,768,0.10601600011189778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,3584,0.00816533300611708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,7168,0.01661777827474806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,12288,0.04598133431540596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,1536,0.02789688772625393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,3072,0.0185217774576611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,51200,28.684461805555557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,8192,0.09484710958268906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,16384,0.02923822071817186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,51200,0.09110044770770603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,2048,0.4450977643330892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,65536,2.8546924591064453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,128,0.006803555621041193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,6144,0.013883555928866068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,3072,0.19681689474317762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,64,0.005704888867007361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,16384,0.6059466467963325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,768,0.012279111478063794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,3584,0.014166222678290473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,8192,0.3087502320607503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,1024,0.004190222256713443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,128,0.006534222099516127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,4096,0.16913510693444145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,8192,2.6824141608344187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,5120,0.018530666828155518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,16384,0.12599289417266846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,32,0.0032746667663256326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,6144,0.16936355166965059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,12288,0.05728622277577718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,32,0.04246222310596042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,128,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,128,0.002700444517864121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,128,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,10240,0.013576000101036496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,12288,0.03305866652064853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,3072,0.03113244308365716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,16384,0.18768355581495497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,4096,0.1781431171629164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,3072,0.06397244665357801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,1536,0.04520444406403435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,768,0.01497866710027059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,4096,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,32,0.005208000126812193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,65536,0.8070018026563855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,5120,0.09700889057583278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,32,0.018778666853904724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,3584,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,16384,0.2816684510972765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,8192,0.18653866979810926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,64,0.012395555774370829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,256,0.009921777579519484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,1024,0.006184000107977126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,128,0.005558222118351195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,768,0.02608977754910787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,5120,0.013023999830087027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,65536,2.649648878309462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,8192,0.027775999572541978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,16384,0.054198222027884595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,6144,0.10996088716718887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,256,0.006568888823191325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,16384,0.18426756064097086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,3072,0.011152889165613385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,16384,0.018331555856598746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,2560,0.006436444405052397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,16384,0.23652267456054688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,1536,0.09858222140206231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,768,0.004919111314747068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,2048,0.025620443953408137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,4096,0.012746666868527731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,2048,0.00701422244310379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,256,0.010552888943089379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,51200,0.773503992292616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,10240,0.27255378829108345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,7168,0.015978667471143935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,7168,0.3408799966176351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,8192,0.0090844440791342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,3584,0.23388444052802193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,2560,0.00700711127784517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,51200,8.955863952636719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,32,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,256,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,5120,0.03415200114250183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,256,0.015855110353893705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,5120,0.02644444505373637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,768,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,10240,0.07479200098249647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,1536,0.008427555362383524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,16384,0.7285635736253527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,2048,0.007059555914666917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,768,0.006145777801672618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,768,0.011112000379297467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,65536,1.2563297483656142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,51200,17.96546088324653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,1024,0.25278843773735893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,7168,0.012176888684431711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,8192,0.3183937867482503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,768,0.0631600022315979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,8192,0.03839644458558824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,5120,0.013426666458447775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,3072,0.009521777431170145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,16384,0.04867466621928745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,51200,0.1143173376719157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,7168,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,2560,0.007506666911972894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,5120,0.017103999853134155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,10240,0.017204445269372728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,32,0.00309955556359556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,16384,0.18807110521528456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,8192,0.013630222115251752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,2048,0.005561777700980504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,512,0.010320888625250923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,12288,0.3895128832923041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,128,0.009617778162161509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,16384,0.023585778143670823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,12288,0.020564445190959506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,768,0.009730666875839233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,10240,2.417672899034288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,65536,0.27280089590284556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,3072,0.0096133334769143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,768,0.01679377754529317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,51200,0.9240453508165148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,3584,0.019658666517999437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,128,0.00313688897424274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,51200,0.6164542304144965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,6144,0.46869511074490017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,64,0.005763555566469829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,7168,0.1172800064086914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,51200,0.10195200310813056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,3584,0.013206221991115145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,768,0.005859555469618903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,65536,0.6184470918443468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,5120,0.3976319895850287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,16384,0.0413680010371738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,51200,0.3227013217078315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,32,0.007261333366235097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,4096,0.027873777680926855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,8192,0.014520888527234396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,1024,0.029807110627492268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,10240,0.21109510792626274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,8192,0.08642755614386664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,1536,0.024732444021436904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,10240,0.1804115507337782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,12288,0.17772800392574736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,4096,0.008580444587601556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,2048,0.01866755551762051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,4096,0.03081511126624213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,64,0.005728000154097875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,768,0.0038248888320393036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,3584,0.009674666656388177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,7168,0.03472266594568888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,1536,0.035551998350355364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,1536,0.28239112430148655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,10240,0.07448177867465548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,64,0.17677688598632812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,32,0.0026311110705137253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,32,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,8192,0.12644978364308676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,256,0.0042257776690853965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,128,0.0072266666425599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,6144,0.13209599918789333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,2048,0.0069439998931354946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,5120,0.021480000681347314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,16384,0.010206222534179688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,768,0.003571555432346132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,128,0.16539466381072998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,6144,0.013167111410035027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,4096,0.019736889335844252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,128,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,16384,0.04537689023547702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,10240,0.024852444728215534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,128,0.00573955559068256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,768,0.018817777434984844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,6144,2.30111206902398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,10240,0.18016089333428276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,5120,0.02071199980047014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,1024,0.04299288988113403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,32,0.008210666477680206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,2048,0.04726577798525492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,3072,0.013232889274756113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,32,0.0123848890264829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,768,0.005611555443869696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,6144,0.020119110743204754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,256,0.0028799999919202593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,768,0.016786666380034555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,512,0.01622577839427524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,6144,0.010072888599501716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,51200,0.27208267317877877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,64,0.002756444530354606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,3072,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,2048,0.011406222151385413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,3072,0.04014755619896783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,51200,0.11656888326009114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,65536,0.648996459113227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,51200,0.10606133275561863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,64,0.007387555307812161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,3584,0.11234488752153184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,2048,0.10011821985244751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,128,0.0027671110712819626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,2560,0.010595555934641095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,2560,0.008873778084913889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,256,0.005591111050711737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,256,0.004435555388530095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,768,0.01902666687965393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,4096,0.05621688895755344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,10240,0.007303111255168915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,768,0.005666666560702854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,64,0.02942488922013177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,5120,0.016004444824324716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,6144,0.015164444843928019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,10240,0.17677066061231825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,16384,0.056207113795810275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,2048,0.009043555292818282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,32,0.0034533333447244433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,7168,0.24169511265224883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,65536,24.483203464084202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,2048,0.011374221907721626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,4096,0.015761777758598328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,1536,0.011628444823953839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,6144,0.021049777666727703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,7168,0.18846844302283394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,512,0.00851999968290329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,2048,0.4811929066975911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,8192,0.5123653411865234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,32,0.008612444831265343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,3072,0.08122666676839192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,6144,0.47219287024603945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,7168,0.030607998371124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,2048,0.07504710886213514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,768,0.012059555285506778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,2560,0.03072355522049798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,1536,0.12183111243777806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,512,0.019069333871205647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,65536,0.16262043846978083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,1536,0.03863111138343811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,2048,0.017143110434214275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,128,0.00281866660548581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,768,0.019549333386951022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,16384,0.08636799785825942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,7168,2.1117350260416665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,65536,0.2556684547000461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,7168,0.010831111007266574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,768,0.005724444571468566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,5120,0.00906311141120063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,1536,0.04100355505943298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,2560,0.018838221828142803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,16384,0.05431644452942742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,8192,0.041257778803507485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,12288,0.028163555595609877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,65536,1.758946630689833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,128,0.005886222339338726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,51200,0.21650755405426025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,1024,0.006574222197135289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,128,0.005096000101831225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,16384,0.019555555449591745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,128,0.013393777940008374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,12288,0.012196444802814059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,2048,0.019303111566437613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,5120,0.009852444132169088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,2048,0.007119110888904995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,128,0.020815999971495736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,128,0.013819555441538492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,65536,0.8888177871704102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,8192,0.007730666961934831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,2048,0.010790222220950656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,128,0.02046133412255181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,2560,0.01716711123784383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,3584,0.11712177594502766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,32,0.008207110895050896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,10240,0.02084622283776601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,4096,0.010505778094132742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,256,0.003133333391613431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,64,0.022663111488024395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,7168,0.02384088933467865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,512,0.006001777946949005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,8192,0.263812435997857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,7168,0.11297155751122369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,3584,0.0654933320151435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,3072,0.010381333529949188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,6144,0.006127999888526068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,2048,0.0053671110007498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,6144,0.023570666710535686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,512,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,51200,0.2731040053897434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,128,0.0052844443255000645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,16384,0.045520888434516064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,512,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,3584,0.22900001207987467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,7168,0.30025243759155273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,1536,0.009518221848540837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,7168,0.010873777998818291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,12288,0.508770677778456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,3584,0.021872000561820135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,256,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,7168,0.0414613319767846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,6144,0.029487109846538965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,51200,21.36465115017361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,256,0.0032355555643637977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,128,0.0029093333416514923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,1024,0.014001778430408902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,65536,0.20570845074123809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,10240,0.018735110759735107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,3584,1.7495270835028753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,256,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,2048,0.011779555843936073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,12288,2.6183395385742188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,4096,0.03191377719243368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,7168,0.015260444747077094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,3584,0.01071911139620675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,6144,0.045054223802354604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,256,0.003718222181002299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,768,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,7168,0.1270888911353217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,768,0.01871555546919505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,7168,0.7421244515313042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,128,0.0030533334033356774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,51200,0.5921351114908854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,3584,0.06758399804433186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,7168,0.0322675539387597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,12288,0.05776355663935343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,768,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,512,0.005583111196756363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,3584,0.30710310406155056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,8192,0.3397502104441325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,3072,0.06407022476196289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,1536,0.009864889085292816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,2560,0.2213724454243978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,12288,0.36978665987650555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,1536,0.004690666579537922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,256,0.0099653336736891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,128,0.005872000008821487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,512,0.0046826667255825466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,128,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,2048,0.008904000123341879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,6144,0.015080889066060385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,64,0.008591999610265097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,768,0.007974222302436829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,32,0.02522666586769952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,3072,0.015164444843928019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,1536,0.0780515538321601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,65536,0.4307084348466661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,65536,1.778656005859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,10240,0.018631110588709515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,3072,0.0097884444726838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,512,0.003830222205983268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,10240,0.05622933308283488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,10240,0.1994160016377767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,256,0.03601333167817857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,8192,0.994090716044108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,65536,1.7758027182685003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,1024,0.004031110968854693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,7168,0.036533332533306546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,64,0.005347555710209741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,65536,0.18568710486094156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,10240,0.09171644184324478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,128,0.004392000122202767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,12288,0.5448862181769477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,65536,0.30017956097920734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,128,0.007320889168315464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,8192,0.020257777637905545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,3584,0.012980444563759698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,256,0.00963644435008367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,4096,0.3733564482794867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,12288,2.1522240108913846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,1536,0.007291555404663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,6144,0.011115555961926779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,64,0.004514666481150521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,3072,0.04299911194377475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,10240,0.026885333988401625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,12288,0.46167198816935223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,1024,0.0425146652592553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,3072,0.01591466698381636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,256,0.1489351060655382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,2048,0.010803555448849997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,8192,0.21091111501057944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,6144,0.038980444272359215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,10240,0.5035715632968479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,6144,0.018001778258217704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,64,0.012469333079126147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,8192,0.007948444121413762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,512,0.009111111362775167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,1536,0.013815999858909182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,65536,2.09205691019694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,3584,0.1469528939988878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,2560,0.01497155593501197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,8192,0.016209777858522203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,1024,0.010334221853150262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,4096,1.7717084884643555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,51200,0.05830133623547024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,51200,0.13460621568891737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,10240,0.026143110460705225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,6144,0.04648177822430929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,10240,0.013365333278973898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,768,0.004095111042261124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,1024,0.02200711104604933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,5120,0.03021777669588725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,10240,0.020450666546821594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,5120,0.08107911215888129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,5120,0.03646844294336107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,768,0.006638222270541721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,3072,0.3598124451107449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,8192,0.03393244412210252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,51200,6.450524648030599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,32,0.006959111326270633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,3584,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,16384,0.03963199920124478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,51200,0.019699555304315355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,3584,0.054895109600490995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,10240,0.019997333486874897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,3584,0.06965244478649564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,6144,0.05440088775422838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,128,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,64,0.0028728888266616394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,5120,0.028885334730148315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,2048,0.012305777933862476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,2048,0.018974221414989896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,5120,0.06022310919231839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,1024,0.010855110983053843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,3072,0.06874577866660224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,4096,0.012853333519564735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,512,0.004629333400064045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,6144,0.0555902189678616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,32,0.0028444443725877335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,1536,0.0070773329999711775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,6144,0.013352889153692456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,3584,0.019344889455371432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,2048,0.005619555711746216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,256,0.013551111022631327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,51200,0.06012977494133843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,1536,0.09035111135906643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,2560,0.006020444548792309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,7168,0.12732088565826416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,5120,0.00999111102686988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,128,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,5120,0.030180444320042927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,51200,1.2859573364257812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,512,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,256,0.03636888994110955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,8192,0.022290666898091633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,3584,0.11930755774180095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,768,0.005736888696750005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,16384,0.7357430987887912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,2560,0.07020178106096056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,32,0.0064293332397937775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,2560,0.018174222773975797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,32,0.0030088888274298776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,2048,0.01000355515215132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,65536,0.43187912305196124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,16384,0.523364437950982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,3072,0.007655111451943715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,2048,0.031643556223975286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,256,0.04308088951640659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,1024,0.014716444744004143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,12288,0.22061067157321504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,65536,0.15600978003607854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,1024,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,10240,0.03261244297027588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,12288,0.054899556769265064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,51200,0.13932889037662083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,256,0.005475555443101459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,64,0.0028320000403457214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,7168,0.021187555458810594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,4096,0.02974489000108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,64,0.002584888910253843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,64,0.005046222358942032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,1024,0.0063031112982167145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,3584,0.012545777691735161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,1536,0.026728888352711994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,1536,0.01072177787621816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,7168,0.4190346664852566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,6144,0.015172445111804538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,3072,0.010976889067225985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,8192,0.019575110740131803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,1024,0.03904177745183309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,65536,0.6703884336683484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,32,0.005611555443869696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,512,0.08047022422154744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,6144,0.37104445033603245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,768,0.019104889697498746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,4096,3.651287078857422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,512,0.0508284436331855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,3072,0.009862222605281407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,4096,0.18503555986616346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,2048,0.006676444576846228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,2560,0.42933421664767796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,7168,0.025295999315049913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,4096,0.03482310970624288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,12288,0.020251555575264823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,3584,0.032786667346954346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,6144,0.009820444716347588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,6144,1.229936917622884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,2560,0.29775823487175834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,65536,2.0013368394639754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,256,0.020596444606781006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,51200,0.30515021748012966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,16384,0.028710222906536523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,1536,0.015379556351237826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,512,0.007292444507280986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,6144,2.32096184624566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,2560,0.15418044726053873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,128,0.0073520003093613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,10240,0.04457599918047587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,3072,0.08266933096779717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,256,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,5120,0.22181067201826307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,2560,0.03510399990611606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,3072,0.05830666753980848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,1536,0.010788444843557147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,4096,0.0751439995235867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,51200,0.8847751087612576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,2048,0.03510399990611606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,6144,0.15937866104973689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,16384,0.9195751614040799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,3584,0.02218933403491974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,7168,0.1366728941599528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,65536,1.1854302088419597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,65536,3.6722399393717446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,1536,0.007340444458855524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,512,0.006339555399285422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,7168,0.6644861963060167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,12288,0.027929776244693335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,65536,2.536543952094184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,768,0.006816888848940532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,256,0.13298488987816706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,6144,0.02572622232966953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,16384,0.024120888776249353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,5120,0.02530044482813941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,2048,0.04822399881150988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,12288,0.5483893288506402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,64,0.00295733329322603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,1024,0.2258711126115587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,65536,1.7494364844428167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,4096,0.18545688523186576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,1024,0.040307554933759905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,2560,0.4305075539482965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,12288,0.07939289013544719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,1536,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,51200,0.9243457582261828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,1536,0.011775111158688864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,512,0.0173662222094006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,5120,0.02344977855682373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,12288,0.10303911235597397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,2560,0.007430222299363878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,12288,0.5324782265557183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,1024,0.004581333448489507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,5120,0.010840000377760993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,16384,0.05158755514356825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,2560,0.009919111099508073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,256,0.0031662223239739737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,1024,0.0058168888919883305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,2560,0.008222222328186035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,7168,0.10478578011194865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,2560,0.053342223167419434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,2048,0.05683110819922554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,1536,0.18880889150831434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,65536,0.8843537966410319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,6144,0.014064888159434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,51200,0.2517706553141276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,3584,0.09474133120642768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,256,0.07367110914654203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,12288,0.025311110748185053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,4096,0.15150311258104113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,10240,0.08621066808700562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,1536,0.05024088753594292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,8192,0.1759706603156196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,4096,0.22011378076341417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,65536,0.07496444384256999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,128,0.002793777734041214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,16384,0.02958044409751892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,5120,0.013519110778967539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,51200,0.834462218814426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,2048,0.011209777659840055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,256,0.004250666747490565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,65536,0.06717866659164429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,4096,0.007858666280905405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,1536,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,8192,0.06554133362240262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,512,0.10333510902192856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,7168,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,512,0.025999110605981614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,128,0.015665777855449252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,5120,0.10103378030988906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,4096,0.006215999937719769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,4096,0.1081119974454244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,10240,0.44223644998338485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,1024,0.017383111847771537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,7168,0.014882667197121514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,256,0.025849777791235182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,1024,0.005211555709441503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,6144,0.03640355666478475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,7168,0.23863021532694498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,1024,0.012615111139085559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,768,0.0034053333931499054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,2048,0.007469333708286285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,1024,0.04725066820780436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,2048,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,1536,0.004516444272465176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,1024,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,16384,0.1415297720167372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,4096,0.010455110834704505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,51200,0.10208000077141656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,512,0.09536977608998616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,12288,0.042865776353412204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,1024,0.057512885994381376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,51200,0.035826666487587824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,3072,0.008357333640257517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,5120,0.2921031051211887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,16384,0.05189866820971171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,2560,0.007972444097201029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,6144,0.2696524461110433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,768,0.003658666792843077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,51200,0.2690968778398302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,8192,0.07911288738250732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,51200,0.10029955705006917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,7168,0.08080266581641303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,5120,0.04068177938461304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,65536,1.6427244610256617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,768,0.005355555564165115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,6144,0.0955386691623264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,4096,1.6766346825493708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,1536,0.010392888552612728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,3584,0.29384978612263996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,64,0.0317884451813168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,4096,0.015124445160230001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,6144,0.01419911119672987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,64,0.03958488835228814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,3584,0.010741333166758219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,8192,0.1449893315633138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,51200,3.596557193332248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,4096,0.05192888776461283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,3072,0.019401777121755812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,32,0.003519111209445529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,6144,4.059830983479817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,512,0.008987555901209513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,1536,0.01941066649225023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,4096,0.0069351109365622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,5120,0.78839111328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,768,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,3584,0.2081528902053833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,3584,0.014705777168273926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,4096,0.014564444621404013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,6144,0.012270222107569376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,65536,0.8161670896742078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,12288,0.15726044442918566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,7168,0.014315555493036905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,2048,0.01128000020980835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,3584,0.010919999745157031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,12288,0.016206221448050607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,512,0.0033111110743549135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,8192,0.12140088611178929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,7168,0.13014400005340576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,256,0.0565777752134535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,128,0.003093333293994268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,6144,0.0070088886552386815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,16384,0.06533510817421807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,256,0.005058666484223472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,256,0.005934222290913264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,16384,11.531889173719618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,128,0.0047102222839991255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,10240,0.07950844367345174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,16384,0.03658577799797058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,3584,0.010593777729405297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,1024,0.009415999882751042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,12288,0.20856444040934244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,6144,0.8813324504428439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,2048,0.008625778059164682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,12288,0.0591848889986674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,16384,0.2631937662760417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,2560,0.007184888753626082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,16384,0.2773635652330187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,7168,0.01593244406912062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,16384,0.11504356066385905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,1024,0.016902221573723685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,1024,0.020485333270496793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,128,0.008051555189821456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,4096,0.01943199998802609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,32,0.0028115556471877625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,3072,0.011158222125636207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,10240,0.015496888094478183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,2048,0.011579555769761404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,1024,0.037913779417673744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,6144,0.08922755718231201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,768,0.33596711688571507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,64,0.011583999627166323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,65536,0.0695306658744812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,3584,0.00889511075284746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,3584,0.010718222293588849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,5120,0.0717137787077162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,10240,6.196592119004991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,32,0.02902844548225403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,8192,0.007376000285148621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,65536,0.3452266587151422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,2048,0.00739733295308219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,65536,0.2034240033891466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,51200,0.02983733349376255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,65536,0.17215022775861952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,768,0.03127822279930115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,6144,0.023636443747414485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,8192,0.031165334913465712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,1024,0.007267555428875818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,768,0.3928568893008762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,32,0.004652444273233414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,10240,0.15433777703179255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,2560,0.1437280045615302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,768,0.0075839997993575195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,5120,0.28489600287543404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,768,0.027970665031009253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,4096,0.07225244575076632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,2560,0.3856293360392253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,128,0.003020444470975134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,2048,0.040947556495666504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,10240,0.8841670884026421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,65536,0.5811093118455675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,2560,0.09905511140823364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,1024,0.015450666348139444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,7168,0.6542204221089681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,12288,0.025081778566042583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,512,0.004265777766704559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,3584,0.09686399830712213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,16384,0.021830222672886316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,512,0.07705866628222995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,1024,0.01219733307758967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,64,0.005741333381997214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,768,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,512,0.021359999974568684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,3584,0.1202053361468845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,768,0.0034862222770849862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,1024,0.0037368887828456033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,768,0.0037235555549462638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,128,0.006213333457708359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,3584,0.00793333351612091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,7168,0.05751199854744805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,1024,0.004321777572234471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,768,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,7168,0.059055997265709766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,512,0.005330666485759947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,2048,0.016524444023768108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,10240,0.02461155586772495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,2048,0.025252444876564875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,7168,0.03273777829276191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,256,0.027318222655190363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,16384,0.5402444203694662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,6144,0.024784000383483037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,2560,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,256,0.013244444297419654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,16384,0.12602489524417454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,32,0.004497777670621872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,7168,0.030540443129009668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,5120,0.029471112622155085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,4096,0.008912000391218398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,1536,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,1536,0.006142222219043308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,10240,6.7716318766276045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,768,0.02126844392882453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,6144,0.04635733366012573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,1536,0.00758755538198683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,65536,0.15609067016177708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,32,0.005049777941571342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,1536,0.02141333288616604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,5120,0.01787911189926995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,3584,0.040720889965693154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,3584,0.01105955574247572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,16384,0.6206915643480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,16384,0.027386666999922857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,8192,0.10879466931025188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,51200,0.1790764464272393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,16384,0.05253155695067512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,1536,0.013214222258991666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,256,0.05298666821585762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,65536,1.3877200020684137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,12288,0.8582666714986166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,32,0.008445333275530074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,8192,0.2501511043972439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,768,0.07244622045093112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,32,0.004971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,6144,0.019458666443824768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,32,0.005809777726729711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,5120,0.042084445556004844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,5120,0.06322489182154338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,10240,3.1115137736002603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,768,0.007502222226725684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,16384,0.9090248743693033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,8192,0.04773688978619046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,2560,0.006843555718660355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,128,0.007450666692521837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,1024,0.21947288513183594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,32,0.012585777375433179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,3584,0.02271555529700385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,12288,0.66029601626926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,2560,0.013181333740552267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,6144,0.12987021605173746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,51200,0.18371289306216768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,10240,0.343178669611613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,2048,0.009837333526876237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,3584,0.26214755905999076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,12288,3.1726515028211804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,16384,0.15189599990844727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,7168,1.0069430669148762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,6144,0.04946044418546888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,7168,0.014533332652515836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,5120,0.02699022160636054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,1024,0.003991111285156674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,1024,0.008247111406591203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,2560,0.016568000117937725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,2560,0.038689778910742864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,4096,0.26158666610717773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,3584,0.18512622515360513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,512,0.003620444486538569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,512,0.05714222457673815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,2048,0.006800888727108638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,16384,0.0365573333369361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,8192,0.03741333219740126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,10240,0.029678222205903795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,3072,0.0192266669538286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,2560,0.00980533328321245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,7168,0.24988089667426217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,1536,0.007656888829337225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,64,0.010264000131024255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,7168,0.03719022207789951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,768,0.012362666428089142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,12288,0.33419378598531085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,3072,0.09028266535864936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,8192,0.015881778465376962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,3584,0.13619732856750488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,51200,1.2353306876288521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,512,0.003968888686762916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,6144,0.018964444597562153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,256,0.004531555705600315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,2048,0.01462311049302419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,6144,0.04123555620511373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,2560,0.04901599884033203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,256,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,8192,0.012404444317022959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,2048,0.24186134338378906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,768,0.009323555562231276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,5120,0.010910222099887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,2048,0.01738133364253574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,3072,0.02169955604606205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,10240,0.015608888533380298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,6144,0.019779556327395968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,2048,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,256,0.010523555179437002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,32,0.002970666728085942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,6144,0.014242667290899487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,256,0.009889778163697984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,3072,0.022749332918061152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,12288,0.13921156194474962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,1536,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,2048,0.06740711132685344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,2560,0.014010666145218743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,128,0.023668444818920557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,10240,0.05036355720625984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,512,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,1536,0.08827466434902614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,768,0.005679111099905438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,128,0.010306666294733683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,1024,0.034903112385008074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,1024,0.1538355615403917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,256,0.047689778937233814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,10240,0.21351467238532174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,1536,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,51200,12.835306803385416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,8192,0.02049511174360911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,512,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,16384,1.7951359219021266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,65536,0.017112889223628573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,1024,0.004671111288997862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,16384,0.02456355591615041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,4096,0.011147555377748279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,64,0.03343644407060411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,128,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,32,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,10240,0.0373271107673645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,16384,0.018431999617152743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,256,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,3584,0.00889155517021815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,5120,0.015404444601800708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,2560,0.049795554743872754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,32,0.002554666664865282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,2560,0.009530666801664565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,51200,4.906755659315321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,65536,0.8574524455600314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,512,0.0184568895234002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,16384,0.021902221772405837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,7168,0.11895644664764404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,64,0.004447111239035924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,64,0.003383999897374047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,8192,0.022434666752815247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,51200,29.621710883246525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,1024,0.04566133353445265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,128,0.0035964445107513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,12288,0.2769991026984321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,2560,0.012400888734393649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,256,0.0032844444115956626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,8192,0.06492800182766385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,1536,0.36812178293863934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,5120,0.00795377790927887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,1024,0.004014222158326043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,1536,0.014075555735164218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,128,0.0029786665820413162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,512,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,7168,0.09777866469489203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,1024,0.019000000423855253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,4096,0.012330667012267642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,64,0.015575110912322998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,256,0.005664888769388199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,1536,0.22042132748497856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,512,0.008714666797055138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,51200,4.7800721062554254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,7168,0.04753511150677999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,65536,0.0657857788933648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,8192,0.14125066333346897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,65536,0.04563733273082309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,16384,0.032524443335003324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,65536,0.6433288786146376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,1536,0.12686488363477919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,12288,0.09617422024408977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,7168,0.023346667488416035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,2560,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,1024,0.04018044471740723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,12288,0.018360889620251127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,7168,0.03134399983617995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,16384,0.054337776369518705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,8192,0.10040711032019721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,2048,0.06440799766116671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,12288,0.01551644504070282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,16384,0.16974222660064697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,2048,0.7644506560431586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,7168,0.011321778098742167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,7168,0.030179556873109605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,6144,0.13055022557576498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,4096,0.009364444348547194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,4096,0.025080889463424683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,128,0.01144977741771274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,512,0.11254578166537815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,51200,14.276645236545138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,6144,0.024636444118287828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,65536,1.0546870761447482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,2048,0.06192977560891045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,3584,0.00979733301533593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,3072,0.008199111455016667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,6144,0.10866310861375596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,256,0.003412444351447953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,3584,0.19488444593217638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,256,0.002842666581273079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,512,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,2560,0.023415111833148535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,51200,0.15303377310434976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,2048,0.15447021855248347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,1536,0.08184978034761217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,7168,0.03720000055101182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,3072,0.006115555349323485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,16384,0.0655226641231113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,12288,0.05190222130881416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,1024,0.24333156479729545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,4096,0.4081466727786594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,256,0.003113777687152227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,3584,0.029948443174362183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,12288,0.07122844457626343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,5120,0.052143110169304736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,6144,0.01830311119556427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,10240,0.03513244456715054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,768,0.7357359992133247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,12288,0.10282399919297959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,51200,5.403113047281901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,256,0.0072266666425599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,7168,0.5158649020724827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,3584,0.500416013929579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,3584,0.015117333994971381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,768,0.02808977829085456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,16384,0.02881244487232632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,2560,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,512,0.010144888526863521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,10240,0.4421786732143826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,6144,0.01882577770286136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,5120,0.009229333036475712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,10240,0.019359999232821994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,1024,0.06255377663506402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,65536,3.851795620388455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,128,0.0069777775141927935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,4096,0.11687466833326553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,16384,0.024119999673631456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,1536,0.009212444225947062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,4096,0.00721866637468338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,128,0.005084444251325396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,1536,0.011236444115638733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,3072,0.08824711375766331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,32,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,12288,0.5412400033738878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,12288,2.0047565036349826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,128,0.012638222012254926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,5120,0.048157334327697754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,65536,0.5855929056803385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,12288,0.019009777241282992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,6144,0.12025333775414361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,5120,0.009907555249002244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,64,0.005551110953092575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,128,0.005299555758635203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,16384,0.028255999088287354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,8192,0.1221546729405721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,2048,0.014627556006113688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,512,0.009462222456932068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,2560,0.009400000174840292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,10240,0.08526400062772964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,65536,0.030925333499908447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,65536,1.1304257710774739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,3072,0.0777733325958252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,32,0.003289777785539627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,512,0.022170666191313002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,3072,0.015440000428093804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,51200,0.06206222375233968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,256,0.01424711114830441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,8192,0.08516977892981635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,768,0.3777040110694037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,2048,0.006214222146405115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,128,0.02156977852185567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,2560,0.008166222108734978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,5120,0.013975110318925647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,256,0.004280000097221798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,16384,0.17279556062486437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,32,0.020202666521072388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,4096,0.20872978369394937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,512,0.021067554752031963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,64,0.006959111326270633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,1536,0.005748444547255834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,16384,0.025078222155570984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,2560,0.020995555652512446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,8192,0.01644977761639489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,7168,0.008778666456540426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,1536,0.0058151111006736755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,7168,0.05198755529191759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,2048,0.024687111377716064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,128,0.02260711126857334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,10240,34.20105658637153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,512,0.004228444563017951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,10240,0.1589911116494073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,256,0.011925333076053195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,2560,0.025755556093321905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,8192,0.09454222520192464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,512,0.00498577786816491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,8192,0.009393778112199571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,2048,0.011967110964987012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,3072,4.250059763590495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,256,0.005004444470008214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,2560,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,128,0.005208000126812193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,64,0.005088888936572605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,2048,0.027607111467255488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,3584,0.014932443698247274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,4096,0.018199111024538677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,51200,0.05464977688259549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,10240,0.26011376910739475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,768,0.037439998653199934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,16384,0.03885689046647813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,768,0.22679644160800508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,32,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,65536,1.7976489596896703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,64,0.03298133280542161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,65536,0.14785956011878118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,512,0.0035422220826148987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,10240,0.02979288829697503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,12288,1.0497360229492188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,256,0.005570666657553778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,5120,0.015027556154463025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,51200,0.2200160026550293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,5120,0.04642311069700453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,12288,0.04527022110091316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,7168,0.030776888132095337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,256,0.005427555491526921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,256,0.017645332548353408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,256,0.03552622265285916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,5120,0.02094488839308421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,128,0.0034471110751231513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,5120,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,64,0.006636444479227066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,32,0.0029519999192820657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,51200,2.8865928649902344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,1536,0.00908088849650489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,51200,0.7019786834716797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,1024,0.2177679936091105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,128,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,3584,0.01382844481203291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,64,0.010211555494202508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,51200,0.3692568937937419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,8192,0.05194488830036587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,8192,0.01388177772363027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,512,0.04431822233729892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,768,0.052742222944895424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,768,0.008942222429646386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,3584,0.011534222298198275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,51200,0.012640888492266336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,768,0.0037066667444176147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,1024,0.022469333476490442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,65536,9.660753038194445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,16384,0.0370506677362654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,128,0.005517333332035277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,2048,0.025097777446111042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,1536,0.4725831349690755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,4096,0.032418668270111084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,3072,0.008567111359702216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,16384,0.0404426654179891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,65536,21.708489312065975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,128,0.005943111247486538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,6144,0.1083644429842631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,5120,0.02105955613984002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,51200,0.4625297652350531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,65536,0.11243200302124023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,1536,0.02046133412255181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,7168,0.018138666947682697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,2048,0.06863555643293592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,12288,0.019479110836982727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,10240,0.32175199190775555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,4096,0.13340800338321263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,2560,0.005956444475385878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,8192,0.07469955417844984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,3072,0.02520977788501316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,3072,0.12796889411078557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,65536,2.7544640435112844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,12288,7.258467780219184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,65536,0.5574737654791938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,51200,1.831994586520725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,64,0.009152889251708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,51200,0.8650124337938098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,10240,0.06026666694217258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,6144,0.17753244770897758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,12288,0.0151582227812873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,3584,0.00870311094654931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,7168,0.06495289007822673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,12288,0.006288888967699475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,5120,0.016275554895401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,12288,0.00608800020482805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,256,0.008023999631404877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,512,0.0032551110618644287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,768,0.006526222245560751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,51200,0.08739911185370551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,1024,0.008273777862389883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,32,0.0025555555605226094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,256,0.003310222178697586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,768,0.049810667832692467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,256,0.06632266441980998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,16384,0.13051644961039224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,8192,0.2284853458404541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,6144,0.038521776596705117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,5120,0.05886222256554497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,16384,0.32046487596299916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,51200,0.7656328943040637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,128,0.0034462221794658234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,3584,0.13112444347805446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,8192,41.800452338324654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,12288,0.05566933419969347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,3072,0.008103999826643202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,7168,0.36103200912475586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,768,0.01074044406414032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,1536,0.013241777817408243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,32,0.003742222156789568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,32,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,51200,0.4018693235185411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,10240,24.206024169921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,3072,0.02206755512290531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,16384,0.05346666773160299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,1024,0.06506399975882636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,10240,0.034543110264672175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,3072,0.014035556051466199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,3072,0.012614222036467658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,4096,0.009666666388511658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,8192,0.007084444165229797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,1536,0.023318222827381555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,7168,0.07227910889519586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,65536,0.06722578075197008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,256,0.005208000126812193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,3584,0.043733331892225474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,3072,0.06014755699369642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,3584,0.6892782317267524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,3584,0.01924533314175076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,2560,0.033238222201665245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,1024,0.005767999837795894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,6144,0.012525333298577202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,8192,0.03309066759215461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,5120,0.23252712355719674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,512,0.007400888535711501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,7168,0.465178648630778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,4096,0.04500888784726461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,768,0.1177191072040134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,5120,0.025860443711280823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,3584,0.009354666703277165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,64,0.003911111089918348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,1536,0.004925333377387789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,128,0.019879110985332064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,128,0.0027066665805048416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,8192,0.0774408910009596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,2560,0.06825511323081122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,12288,0.18244532744089761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,4096,0.00582755563987626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,6144,0.014256000518798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,64,0.020005333754751418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,128,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,10240,0.04424533247947693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,51200,2.5335110558403864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,4096,0.019457777341206867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,10240,0.035559111171298556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,12288,0.030925333499908447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,2048,0.46954578823513454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,51200,0.10140622324413723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,256,0.02256800068749322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,32,0.007391110890441471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,65536,1.5912435319688585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,128,0.004986666556861666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,2560,0.2719493442111545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,1536,0.008953777452309927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,3072,0.006796444455782573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,32,0.003027555636233754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,512,0.0058808889653947614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,6144,0.029302222861184016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,12288,0.03527288966708713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,1024,0.006638222270541721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,1536,0.014608888162506951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,1024,0.01867377758026123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,10240,0.019116444720162284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,10240,0.027505778604083594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,256,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,8192,0.05374044511053297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,16384,0.011688888900809817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,51200,0.22780977355109322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,768,0.003588444242874781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,128,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,1536,0.011301333705584208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,256,0.005372444374693765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,768,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,65536,0.06755466593636407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,12288,0.012778666284349231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,6144,0.023220444718996685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,6144,0.08469688892364502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,128,0.0041555555330382455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,512,0.00507733349998792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,128,0.0026444445053736367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,1536,0.07237866852018568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,1536,0.0064479998416370815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,1536,0.008929777476522658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,256,0.09427644146813287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,3584,0.012484444512261285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,1024,0.004091555459631814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,10240,0.013439999686347114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,32,0.008360000120268928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,1024,0.08475910955005223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,3072,0.009180444810125563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,7168,0.01978133287694719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,5120,0.48374578687879777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,4096,0.01939022209909227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,16384,0.7050737804836698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,1536,0.07806133561664157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,65536,0.06600444184409247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,1536,0.009389333426952362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,256,0.005537777725193236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,4096,0.01754488878779941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,65536,5.719117482503255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,2048,0.1170568863550822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,128,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,51200,0.051691555314593844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,12288,0.015145777000321282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,2560,0.05197422371970283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,1024,0.04442755712403191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,512,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,3584,0.007439110842016008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,8192,0.04855377806557549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,3072,1.2641288969251845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,5120,0.012404444317022959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,2048,0.008234666453467475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,10240,0.6008079846700033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,51200,0.18376444445716011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,5120,0.025482666161325242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,3072,0.017272000511487324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,51200,0.058238221539391406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,65536,0.04071288969781663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,256,0.007658667034573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,3584,0.011785777906576792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,8192,0.505475574069553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,12288,0.0877422226799859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,2048,0.016528889536857605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,10240,0.019010666343900893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,1536,0.029875556627909344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,32,0.0035208890007601846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,512,0.02532711128393809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,1536,0.0063795554969045855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,768,0.10838933123482598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,3584,0.017796445224020217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,10240,0.2226462099287245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,2048,0.14098310470581055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,2048,0.023331556055280898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,3072,0.016693333784739178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,128,0.0030168888883458245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,5120,0.018073777357737224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,768,0.011759999725553723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,5120,0.030598223209381104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,5120,0.008063111040327284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,51200,3.0844978756374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,7168,0.026467555099063452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,2048,0.0069671107663048645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,256,0.004363555461168289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,51200,1.9118720160590277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,2560,0.08542755577299331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,512,0.011540444360838996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,51200,0.22450666957431367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,7168,0.150764438841078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,768,0.01516177753607432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,51200,13.55737559000651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,64,0.006519111080302133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,512,0.053335110346476235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,64,0.03204710947142707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,2048,0.011122666299343109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,256,0.0358551111486223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,2560,0.018763555420769584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,128,0.011642666326628791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,128,0.0029022221763928733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,1536,0.0058408888677755995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,7168,0.013590222431553734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,5120,0.018615111708641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,16384,0.04704088966051737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,65536,0.03922933340072632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,8192,0.1521031061808268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,16384,0.006196444647179709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,7168,0.08283378018273248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,256,0.0033200000309281876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,3584,0.0664106673664517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,10240,0.018570666511853535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,1024,0.005015111217896144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,3584,24.035695393880207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,65536,0.14833866225348577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,12288,0.03555999861823188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,3072,0.009833777944246927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,3584,0.00906311141120063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,5120,0.02327022287580702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,5120,0.02055022286044227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,16384,0.36269511116875547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,2048,0.2167066600587633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,64,0.0074417781498697065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,1536,0.011230222052998014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,4096,0.28285243776109487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,2048,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,256,0.0041626666982968645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,1536,0.006663110935025745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,128,0.004443555656406614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,1024,0.03507377703984579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,1024,0.007937777373525832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,12288,0.04166933231883579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,1024,0.006114666660626729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,5120,0.01584533353646596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,65536,42.423441569010414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,3072,0.1517066690656874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,32,0.0031759999692440033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,5120,0.5251395437452528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,65536,0.08116088973151313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,5120,0.031000001562966242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,768,0.008398222426573435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,3072,0.020830222302012973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,7168,0.030377778742048476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,64,0.03308444552951389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,128,0.007311999797821045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,3072,0.1090008881356981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,10240,0.059566219647725425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,12288,0.2827440102895101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,51200,0.6717822286817762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,4096,0.009344889058007134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,10240,0.10207288795047337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,128,0.0053431110249625305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,5120,0.02088800072669983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,7168,0.018790221876568265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,256,0.010814222196737925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,10240,0.10284622510274251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,1024,0.006573333508438534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,3584,0.040447112586763166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,7168,0.02923911147647434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,65536,0.437824010848999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,12288,0.08585688802931045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,4096,0.04273777869012621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,65536,3.026235580444336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,12288,0.3289795451694065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,12288,0.15343288580576578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,2560,0.01145155562294854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,768,0.10034400224685669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,2048,0.03562577896647983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,16384,0.04483555422888862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,1536,0.009376000199053023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,6144,0.024903999434577093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,3584,0.01514933341079288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,64,0.0027155555370781156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,51200,0.04890755481190152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,8192,0.032960888412263654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,768,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,1536,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,2048,0.06352266338136461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,1536,0.0046773333516385825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,5120,0.01131199962562985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,256,0.011083555718262991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,2560,0.07429866658316718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,4096,0.2762311034732395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,256,0.0374951097700331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,1024,0.006416888700591193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,1536,0.1677884393268161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,65536,0.07662044631110297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,5120,0.028421332438786823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,16384,0.21319999959733751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,3584,0.037267555793126426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,7168,0.06674844688839383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,1024,0.004047111090686587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,4096,0.011791110866599612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,32,0.0034008889148632684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,7168,0.16604355971018472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,512,0.0033733333564466904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,2048,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,65536,0.25924889246622723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,768,0.0047191112405723995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,8192,0.006130666782458623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,12288,1.202223989698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,16384,0.05536444319619072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,8192,0.16210577223036024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,4096,0.07714844412273832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,65536,0.5001769065856934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,6144,0.009954666925801171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,64,0.005323555734422471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,32,0.008591999610265097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,51200,0.07029777765274048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,16384,3.569597456190321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,4096,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,3072,0.05753778086768257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,128,0.005200888961553574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,512,0.0053004444473319584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,16384,0.02283199959331089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,1024,0.03662755423121982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,65536,0.22452889548407662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,16384,0.27030399110582143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,4096,0.01144444445768992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,4096,0.09968533118565877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,128,0.004262222184075249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,64,0.005095110999213325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,6144,0.3902328809102376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,1536,0.7489902178446451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,16384,0.05257066753175524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,256,0.07727911074956258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,3584,0.09680977794859146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,768,0.04692000150680542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,10240,0.05001777741644117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,6144,0.09688978062735663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,4096,0.014345778359307183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,16384,0.007941332956155142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,16384,0.02688088847531213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,51200,0.07009866502549914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,16384,0.10747644636366104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,768,0.053130666414896645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,16384,0.04015644391377767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,12288,0.03071911136309306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,1536,0.09946488671832615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,1024,0.006704000135262807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,1024,0.020928000410397846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,3584,0.03178222311867608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,8192,0.6368204222785102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,3072,0.011115555961926779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,7168,0.008160889148712158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,128,0.0028826666788922418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,2560,0.017973333597183228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,2048,0.01827911039193471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,768,0.02354755500952403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,1024,0.0047768888374169665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,768,0.005635555419656966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,256,0.013783999615245394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,3072,0.01033866653839747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,3584,2.441759957207574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,512,0.02606666584809621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,128,0.006167111297448476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,2560,0.10639288690355088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,8192,0.03296444482273526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,7168,0.309441778394911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,6144,0.04170133339034187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,12288,0.03708533445994059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,2560,0.08513689041137695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,5120,0.006881777611043718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,16384,1.0893760257297092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,2560,0.00740533322095871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,2048,0.030823998981051978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,1536,0.15652444627549914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,10240,0.21819821993509927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,768,0.005535111245181825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,2560,0.007167999943097432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,512,0.030720889568328857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,3072,0.6967128647698296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,4096,0.09014933639102513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,51200,0.9155431323581271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,1024,0.09802133507198757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,2048,0.05277422070503235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,64,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,12288,0.30913954310946995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,6144,0.011139555937714048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,6144,0.6209608713785807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,16384,0.009002666506502364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,51200,0.06948444578382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,3072,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,16384,0.16579466395907932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,3584,1.4941342671712239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,5120,0.2217235565185547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,512,0.004078222231732475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,51200,0.6350088649325901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,5120,0.03518844313091702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,3584,0.010286221901575724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,64,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,512,0.005295111073387994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,7168,0.3264275656806098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,5120,0.07202933231989543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,128,0.007798222204049428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,64,0.005327111317051782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,3072,0.06568977567884657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,1536,0.007261333366235097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,2048,0.016662221815851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,10240,5.0035349527994795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,51200,0.029915554655922785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,65536,0.21641600131988525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,8192,0.014923555983437432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,768,0.19422933790418836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,256,0.003714666598372989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,2560,0.021447999609841242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,8192,0.027448000179396734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,65536,0.25013687875535756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,10240,0.029603554142846003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,10240,0.07491555478837755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,2048,0.10659288697772556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,3584,0.007983999947706858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,16384,0.051550222767723926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,12288,42.19085354275173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,256,0.012114666402339935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,3072,0.6318657663133409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,3072,0.08758044242858887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,3072,0.18201333946651885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,1024,0.0047528888616296975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,3072,0.009846222069528367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,128,0.007127111156781514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,12288,0.007107555866241455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,16384,0.024698666400379602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,256,0.02092622220516205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,51200,0.04965066578653124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,10240,0.06718133555518256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,6144,0.028598222467634413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,51200,0.11776088343726264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,2048,0.007855110698276097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,1536,0.006395555618736479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,64,0.16776977645026314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,51200,1.8694915771484375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,256,0.004274666723277834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,4096,0.19592177867889404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,2560,0.007214222517278459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,64,0.006072000082996156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,5120,0.010392888552612728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,7168,0.03088888857099745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,2048,0.010769777827792697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,512,0.004649777793222004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,6144,0.280295107099745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,3584,0.11810044447580974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,2560,0.16403733359442815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,1536,0.0047253333032131195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,4096,0.21815200646718344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,16384,0.05200622147983975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,6144,0.05366933345794678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,51200,0.0610719985432095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,6144,0.02278222143650055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,51200,0.028014222780863445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,6144,0.011884444289737277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,5120,0.02345688806639777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,4096,0.20102132691277397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,64,0.006049777898523543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,2048,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,8192,0.006830222076839871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,7168,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,10240,0.02346399923165639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,10240,0.11555733945634629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,4096,0.10269866387049358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,32,0.012862222062216865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,7168,0.014225777652528552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,1536,0.052777777115503945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,64,0.186097780863444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,65536,0.2643520037333171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,65536,0.03703644540574815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,256,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,512,0.34573955006069607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,64,0.20925245020124647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,5120,0.017078222499953378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,10240,0.08025599850548638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,3072,0.41623110241360134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,512,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,2048,0.0074720001882976955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,8192,0.21024710602230498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,65536,0.9476711485120984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,10240,0.03235910998450385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,256,0.01755288905567593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,65536,0.2605484326680501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,7168,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,256,0.02342044479317135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,65536,19.410152859157986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,5120,0.10682133171293472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,2048,0.06122933493720161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,128,0.0034453332838084963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,3584,0.12169067064921062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,32,0.005493333356248007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,1024,0.08631200260586208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,2560,0.11793066395653619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,51200,0.8465520011054145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,7168,0.7041884528266059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,3584,0.327294217215644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,8192,0.06547200017505221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,2048,0.03788444399833679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,7168,0.36690934499104816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,3584,0.019913777709007263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,1024,0.015944000747468736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,65536,0.09407555394702488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,1024,0.0038568890757030914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,10240,0.14643643962012395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,12288,0.019865777757432725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,512,0.0068400001360310455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,7168,0.23809599876403809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,10240,0.05608977874120077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,1536,0.08239289124806722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,5120,0.44191379017300075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,1024,0.012272888587580787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,2048,0.006934222247865465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,4096,0.010415111151006486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,3584,0.239245335261027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,128,0.007461333440409766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,1024,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,512,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,128,0.02630133264594608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,5120,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,2560,0.010307555397351583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,2560,0.012109333442317115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,64,0.002467555511328909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,128,0.0028702221396896574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,6144,0.019169777631759644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,7168,0.019983111156357657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,64,0.0026551110463009942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,3584,0.008952000074916417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,8192,0.09501866499582927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,512,0.005404444618357553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,2048,0.005382222019963794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,12288,0.10833066701889038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,8192,0.030790223015679255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,256,0.0029582221888833572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,32,0.006552000012662675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,10240,0.03131466772821214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,512,0.04480088750521342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,3584,0.014906666345066495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,3072,0.00851733320289188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,768,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,16384,0.006385777973466449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,3072,0.013980444934633044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,65536,0.10500088665220474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,512,0.022728888524903193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,16384,0.04130666785769992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,10240,0.01350400017367469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,65536,0.5804106924268935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,3072,0.0063324446479479475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,6144,0.05887111028035482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,7168,0.05216888917817009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,4096,0.18220089541541207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,7168,0.024121777878867254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,2048,0.16010488404168025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,8192,0.2151635620329115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,1536,0.0063697778516345555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,10240,0.014579556054539151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,51200,2.760328928629557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,12288,0.0226275556617313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,512,0.02053600052992503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,768,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,256,0.005340444544951121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,7168,0.033795555432637535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,6144,0.09009333451588948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,7168,0.15406489372253418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,3072,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,1024,0.01107377807299296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,2048,0.02199555602338579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,51200,0.2496862146589491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,51200,0.05257066753175524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,7168,0.01054666688044866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,768,0.004418666578001446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,8192,0.11786044968499078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,1536,0.018378666705555387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,51200,0.053563555081685386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,4096,0.008800889054934183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,16384,0.11734755833943684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,5120,0.08518399794896443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,32,0.0028311111446883944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,64,0.0041955556306574075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,256,0.07402311431037055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,1536,0.21188088258107504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,1024,0.009859555297427708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,32,0.0026586666289303037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,7168,0.11869422594706218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,65536,1.5343572828504775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,16384,0.061781333552466504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,256,0.023758222659428913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,3072,0.008802666432327693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,4096,0.009279111193286048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,768,0.014372444815105863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,3584,0.014373333917723762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,256,0.0036755556033717263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,8192,0.00701422244310379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,512,0.0032702222880389956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,6144,0.026458667384253606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,51200,0.057348443402184375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,5120,0.014096889230940076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,2048,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,51200,1.964779535929362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,128,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,256,0.008967111508051554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,7168,0.04803911182615492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,3072,0.19401689370473227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,4096,0.009846222069528367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,128,0.0047031111187405055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,12288,0.03747911254564921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,1024,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,768,0.004554666578769684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,512,0.01681422193845113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,16384,0.1515173382229275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,512,0.04519644379615784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,3584,0.017993777990341187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,256,0.011167111496130625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,768,0.048113776577843555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,1536,0.07001155614852905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,2048,0.007615110940403408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,10240,0.21336266729566786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,10240,0.05668533510631985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,10240,0.025018667181332905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,3072,0.008531555533409119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,1536,0.024521778027216595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,512,0.015432000160217285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,3072,0.00834222220712238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,2560,0.021632888250880774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,768,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,2048,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,2560,0.007134222322040134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,32,0.005958222266700532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,3072,0.1727422210905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,4096,0.0064177778032090925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,1536,0.03362222181426154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,3072,0.007640889121426477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,4096,0.06394577688641019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,768,0.18425244755215117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,768,0.11598132716284858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,5120,0.037371555964152016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,5120,0.10507910781436497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,4096,0.058800889386071094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,3072,0.007804444266690149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,1536,0.011737777955002256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,4096,0.011038222246699862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,10240,0.25935199525621205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,3072,0.031621333625581526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,10240,0.07006488906012641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,3072,0.08036710818608601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,3584,0.010987555815113915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,1024,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,256,0.003113777687152227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,8192,0.009277332988050249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,768,0.04617422156863742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,7168,0.011005333728260465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,32,0.006826666494210561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,8192,0.4504106839497884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,4096,0.023927110764715407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,3584,0.013501333693663279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,12288,0.16144622696770564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,128,0.0034266666819651923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,3584,0.12538133727179632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,512,0.0035031110876136353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,12288,0.019201777047581144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,12288,1.0430079566107857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,6144,0.01276177747382058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,4096,0.04566488994492424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,1024,0.0309804446167416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,65536,0.08380711078643799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,3072,0.007640889121426477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,65536,0.21918043825361463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,7168,0.022659555077552795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,65536,0.2681573232014974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,12288,0.01975200076897939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,3584,0.024967110819286768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,256,0.004253333227501975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,8192,0.01868799991077847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,1536,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,1536,0.024626667300860088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,5120,0.23879732025994194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,2048,0.03479555580351088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,6144,0.021035555336210463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,1024,0.07433688640594482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,3584,0.013222222526868185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,512,0.05190222130881416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,2048,0.01644444465637207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,16384,0.21454755465189615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,768,0.024243555135197107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,5120,0.025847999585999384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,4096,0.03270577722125583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,5120,0.15447999371422663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,2560,0.0749102234840393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,2560,0.019115555617544387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,5120,0.12483644485473633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,128,0.006899555524190267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,16384,0.105003555615743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,768,0.17366222540537515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,768,0.005931555396980709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,65536,0.08537777927186753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,51200,2.864638222588433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,2048,0.05726222197214762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,5120,0.1694906685087416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,7168,0.16713599363962808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,2560,0.00792711145348019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,4096,0.012244444754388599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,8192,0.014053333136770459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,3584,0.008718222379684448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,6144,0.15771378411187068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,128,0.00868711123863856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,16384,0.021193777521451313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,2560,0.60262664159139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,12288,32.099955240885414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,768,0.018262222409248352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,2048,0.0074622225430276655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,768,0.004270222038030624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,128,0.022312000393867493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,1536,0.009117333425415887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,512,0.008625778059164682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,5120,0.020253333780500624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,768,0.0036391110883818734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,256,0.004927111168702443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,512,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,3584,0.049288001325395375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,2560,0.05178577701250712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,32,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,64,0.006569777925809224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,1536,0.13033955627017552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,65536,0.9817866219414605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,12288,0.0883786678314209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,2048,0.01295644458797243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,64,0.3394755522410075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,128,0.0030337776988744736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,5120,0.01165511127975252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,6144,0.01387111097574234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,3584,0.028734223710166082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,3584,0.2440755632188585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,65536,0.014800889624489678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,6144,0.15481600496504042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,10240,0.014130666851997375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,3072,0.012649777862760754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,32,0.022426666484938726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,10240,0.14380621910095215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,32,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,4096,0.006699555450015598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,51200,0.6776329146491157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,128,0.004571555389298333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,512,0.03151999910672506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,51200,0.29289065466986763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,65536,10.763841417100695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,51200,1.0828018188476562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,4096,0.04906310968928867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,16384,0.007255111303594377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,2560,0.2107306718826294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,3072,0.03254133462905884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,768,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,6144,0.0111191107167138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,1536,0.016301333904266357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,3584,0.09509599871105617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,5120,0.4640648629930284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,1024,0.02847466535038418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,7168,0.8729564348856608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,65536,0.010486221975750394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,1024,0.010279110736317104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,256,0.03725511166784499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,2048,0.21831377347310385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,2560,0.1368231111102634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,10240,0.3936951160430908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,2560,0.006340444501903322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,6144,0.02088711162408193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,2560,0.007081777685218387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,8192,0.026657778355810378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,64,0.010562666588359408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,3072,0.0671022203233507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,2048,0.005471999860472149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,3072,0.011168888873524137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,768,0.007782222496138678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,10240,0.016202666693263583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,1024,0.006599999964237213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,10240,0.025184889634450276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,4096,0.0099653336736891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,10240,0.031975110371907554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,5120,0.20148532920413545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,768,0.013029333617952136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,7168,0.019814221395386588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,7168,0.01772711177666982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,51200,0.42501423094007706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,2560,0.03413688805368211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,32,0.012628444366984896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,64,0.127402663230896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,65536,0.01314755529165268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,128,0.003321777822242843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,6144,0.10990222295125325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,16384,0.39079467455546063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,3584,0.009776888622177972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,12288,0.026741334133678015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,128,0.019375110665957134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,7168,0.12274666627248128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,128,0.051967998345692955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,5120,0.030625777112113103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,7168,0.046442667643229164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,128,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,256,0.0030542222989930045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,3584,0.05673778057098389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,6144,0.3020453453063965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,512,0.003478222423129611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,2560,0.007557333343558841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,1024,0.008599999878141616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,16384,0.04489511251449585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,1024,0.048319111267725624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,51200,23.911298963758682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,4096,0.10522400008307563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,3584,0.051273776425255664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,16384,0.018553778529167175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,64,0.007844444778230455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,10240,0.31437600983513725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,6144,0.031007111072540283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,16384,0.7916737662421333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,128,0.004275555411974589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,10240,0.03777155611250136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,16384,0.7539457745022244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,16384,0.05166844526926676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,6144,0.023323555787404377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,32,0.003076444483465619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,4096,0.025191111697090998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,768,0.006320000108745363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,1536,0.06100533405939738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,768,0.02176622218555874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,7168,0.0069448889957533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,256,0.007110222346252865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,2048,0.5165608723958334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,2560,0.019116444720162284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,2048,0.037348445918824934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,10240,0.029142220815022785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,8192,0.024517332514127094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,64,0.00517155561182234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,12288,0.018039999736679923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,768,0.01980799933274587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,1536,0.013976888524161445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,32,0.0030951110853089225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,6144,0.12115821573469375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,12288,0.05592177973853218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,51200,0.44070667690700954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,10240,0.027993778387705486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,65536,3.376097785101997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,16384,41.451314290364586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,1024,0.016367110941145156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,8192,0.025850666893853083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,1024,0.02793333265516493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,1536,0.004690666579537922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,7168,0.7265751096937391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,5120,0.017895999881956313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,6144,0.13014754984113905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,64,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,3072,0.12174933486514622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,10240,0.07730044258965386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,10240,2.970963584052192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,12288,13.523004319932726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,16384,0.043770667579438954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,10240,0.33249688148498535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,16384,0.06029778056674533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,256,0.014145778285132514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,768,0.006215111249023014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,8192,0.011704888608720569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,512,0.0036888888312710654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,2560,0.06913066572613187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,768,0.02069244450993008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,3584,0.18189333544837105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,2560,0.09978933466805352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,8192,0.015101333459218344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,65536,0.06117866436640421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,7168,0.10578844282362197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,2048,0.20065600342220732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,32,0.00509333320789867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,8192,0.022914666268560622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,10240,0.04292444388071696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,16384,0.061965333090888135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,4096,0.00962933318482505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,1024,0.0693253344959683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,7168,0.012605333493815528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,2560,0.08716977967156304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,512,0.038760887251959905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,512,0.00628355559375551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,32,0.005172444300519096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,7168,0.46467113494873047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,65536,0.2577013439602322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,512,0.004431999805900785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,64,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,128,0.11040888892279731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,1024,0.020766221814685397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,10240,0.26520178053114146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,4096,0.10416622294320001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,1536,0.037608888414171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,16384,0.029351999362309773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,32,0.007547555698288812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,2048,0.014118222726715935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,256,0.00867555538813273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,512,0.021205332544114854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,5120,6.206314510769314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,768,0.007634667058785756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,16384,0.09537333250045776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,3584,0.019800888167487252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,7168,0.020608888732062448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,5120,0.011081777513027191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,7168,0.10937777492735122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,51200,0.9380559921264648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,768,0.011056000159846412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,1536,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,3072,0.04865333437919617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,64,0.005574222240183089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,12288,0.16423911518520778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,768,0.018191110756662156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,512,0.0032755554550223877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,2048,0.02896089024013943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,7168,0.00646666685740153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,768,0.03146044413248698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,51200,2.5228453742133246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,512,0.005462222215202119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,7168,0.1400337749057346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,3072,0.008658666577604082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,10240,0.015396444333924187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,2048,0.005037333402368757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,4096,0.5111173523796929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,51200,0.8766844537523059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,1536,0.04124266571468777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,51200,0.09331288602617051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,1024,0.004010666575696734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,51200,0.1853848828209771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,1024,0.003984000119898054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,1024,0.014512888259357877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,1024,0.005619555711746216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,6144,0.07051110929912992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,8192,0.05416088965204027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,2560,0.012335110869672565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,10240,0.016498666670587327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,1024,0.31018132633633083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,3584,0.05378489030732048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,32,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,5120,0.11547644933064778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,128,0.030326220724317763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,1536,0.012295111185974546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,512,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,2048,0.03776444329155816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,32,0.004920000003443824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,3072,0.007337777978844113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,12288,0.17821421888139513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,6144,0.016773333152135212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,4096,0.016536000702116225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,16384,1.973897722032335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,12288,0.029616889026429918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,7168,0.31172000037299263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,1536,0.17580710517035592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,6144,0.009565333525339762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,2560,0.03551200032234192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,4096,0.011416888899273343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,5120,0.034959110948774554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,3584,0.006353777729802662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,4096,0.015561777684423657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,256,0.015455999308162265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,5120,0.006731555693679386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,1536,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,512,0.003228444399105178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,7168,0.010830221904648675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,10240,0.018650665879249573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,128,0.006966222491529252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,256,0.01014755583471722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,2048,0.014595554934607612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,3584,0.008340444829728868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,64,0.005260444349712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,10240,0.01220355596807268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,7168,0.08870933453241985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,2048,0.04463644491301643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,2560,0.02340000040001339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,3072,0.0673324465751648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,3072,0.008614222208658854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,128,0.0029235554652081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,3072,0.012655999925401477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,2560,3.6636479695638022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,3072,0.008266666697131263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,3584,0.01294844432009591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,3072,0.009658666948477427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,3584,0.01533066729704539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,128,0.004620444443490771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,4096,0.007428444094128079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,768,0.021556443638271753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,51200,0.2714853286743164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,65536,0.21124888790978325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,65536,31.221808539496525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,64,0.0034755555291970572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,64,0.0032986667421129015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,128,0.052728000614378184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,10240,0.027708444330427382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,256,0.017749332719379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,256,0.011706666813956367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,2560,0.008437333007653555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,5120,0.17680622471703422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,1024,0.194304002655877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,3584,0.006976889239417181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,3072,0.06198044617970785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,8192,0.03732622332043118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,8192,0.027655111418830022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,256,0.024075556132528517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,2048,0.08713333474265204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,12288,0.3303164376152886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,16384,0.019890666007995605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,65536,8.14840359157986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,4096,0.015071110592948066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,8192,0.023056889573733013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,12288,0.1403075589074029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,51200,0.12039732933044434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,12288,0.11182310846116807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,7168,0.019104000594880845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,1024,0.012609778179062737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,256,0.008231110870838165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,128,0.07180444399515788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,3072,0.15108799934387207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,512,0.07021600008010864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,2560,0.013590222431553734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,12288,0.018328888548745047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,3584,0.0847617785135905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,10240,4.423131730821398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,4096,0.061168001757727734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,1536,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,512,0.006240889016124938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,2048,0.006753777878151999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,65536,3.6926311916775174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,1024,0.008061333662933773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,512,0.00397777764333619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,128,0.02500533395343357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,1536,0.025012445118692186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,1024,0.004849777867396672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,16384,0.1282800038655599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,6144,0.7465199894375272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,32,0.0038044444388813446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,1024,0.07526310947206286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,1024,0.011742221812407175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,768,0.007196444604131911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,3584,0.008082666330867344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,128,0.005366222312053044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,64,0.0063822223908371395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,2048,0.006272000157170826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,2560,0.012312889099121094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,2048,0.05929155482186211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,512,0.004294222427739037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,2560,0.012618666721714867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,65536,0.1341466638776991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,1536,0.004984000076850255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,12288,1.2150239944458008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,2560,0.027090667022599116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,32,0.0034311111602518293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,1536,0.005767111149099138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,6144,0.010215111076831818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,3072,0.008013333711359236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,1536,0.0521360006597307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,16384,0.5231555302937826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,7168,0.049476444721221924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,256,0.008192000289758047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,512,0.007019555403126611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,65536,0.8829351001315646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,10240,0.2523182233174642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,1024,0.006632888896597757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,2560,0.08505333132214016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,64,0.007150222029950883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,3072,0.00979555563794242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,4096,0.00871733327706655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,64,0.0069208890199661255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,3072,0.013236444857385425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,4096,0.21557954947153726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,6144,0.055883553293016225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,4096,0.013783999615245394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,16384,0.025642666551801894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,768,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,16384,0.5254097514682347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,16384,0.5735271241929796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,10240,0.09419288900163438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,1024,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,7168,0.03387199838956197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,5120,0.01589866644806332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,4096,2.9741831885443797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,4096,0.014336888988812765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,1536,0.03983111182848612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,512,0.006090666684839461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,3072,0.011513777905040316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,51200,0.11387111081017388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,12288,0.03659644391801622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,7168,0.07718577649858263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,64,0.002776888923512565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,3072,0.1447991132736206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,256,0.024404444628291663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,128,0.00508355556262864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,16384,0.03148533238304986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,768,0.007684444387753804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,5120,0.07364355855517916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,128,0.018301332990328472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,65536,0.5341279771592882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,10240,0.2778711054060194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,4096,0.009413333402739631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,2560,0.066648887263404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,10240,0.023931556277804907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,256,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,7168,0.057020445664723717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,4096,0.017934223016103108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,65536,0.14550222290886775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,2560,0.05860000186496311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,7168,0.02888266576661004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,65536,0.07591911156972249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,128,0.002777777819169892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,2560,0.023473777704768713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,1024,0.009503111243247986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,1024,0.04851466748449537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,768,0.01756266587310367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,1536,0.7030791176689996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,16384,0.3113528887430827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,8192,0.07507733503977458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,512,0.007873777714040544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,6144,0.024322667055659827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,6144,0.017464000317785475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,3584,0.010078222387366826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,768,0.009316444396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,256,0.017488888568348356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,1536,0.08503378099865383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,16384,0.06029155519273546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,32,0.002532444480392668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,12288,0.34514665603637695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,768,0.006193777753247156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,65536,0.06753600305981107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,65536,0.21889955467647979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,65536,1.797732459174262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,4096,0.048876444498697914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,1024,0.03641511003176371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,4096,0.8860070970323352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,12288,0.033927109506395124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,4096,0.028385778268178303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,4096,0.016365332735909354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,65536,0.9098729027642144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,768,0.006073777874310811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,10240,0.03465422325664096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,6144,0.2843902111053467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,7168,0.06307555569542779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,768,0.012676444318559436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,512,0.016726222303178575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,3072,0.012886222038004132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,3072,0.04745777779155307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,16384,0.10761155684789021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,2560,0.007994666695594788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,5120,0.024085332949956257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,12288,0.3037697739071316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,10240,0.01886577738655938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,128,0.017996443642510306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,7168,0.010805333654085795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,256,0.014942222171359591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,512,0.07710933023028903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,5120,0.017895999881956313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,3584,0.07659288909700182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,8192,0.11459022098117405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,64,0.011799111134476133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,2048,0.02997066577275594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,2048,0.010890666809346942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,3072,0.054971555868784584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,2048,0.199837327003479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,8192,0.7556399769253201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,2560,0.017473777135213215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,65536,0.9544284608629016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,1024,0.0064479998416370815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,1024,0.011567999919255575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,3584,0.037581334511439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,2048,0.01094133324093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,2560,0.011890666352377998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,2048,0.007843555675612556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,12288,0.11065244674682617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,10240,0.743468443552653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,256,0.04332888788647122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,4096,0.028735111157099407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,1024,0.025430222352345783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,32,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,5120,0.1378426684273614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,65536,0.07096888621648152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,8192,0.030642665094799463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,512,0.005271111097600725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,4096,0.039119998613993325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,256,0.007689778175618913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,768,0.004508444418509801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,8192,0.018775110443433125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,32,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,16384,3.980645497639974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,4096,0.14325777689615884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,8192,0.025759999950726826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,10240,0.015761777758598328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,768,0.004829333474238713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,2560,0.021475555168257818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,2560,0.008058666355080074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,65536,1.0217706892225478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,64,0.0034906665484110513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,768,0.01836266616980235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,7168,0.04645511176851061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,5120,0.01019377758105596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,12288,0.15320266617669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,6144,0.018975110517607797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,2048,0.06855377886030409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,2560,0.05450133482615153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,51200,0.8846053547329373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,5120,0.08708710802925958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,16384,0.24198667208353677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,512,0.04226933254135979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,1536,0.014614222778214348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,16384,5.196606106228299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,8192,0.021239999267790053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,7168,0.1669724384943644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,32,0.006080889039569431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,12288,0.522938675350613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,1536,0.23807999822828504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,12288,0.21507910887400308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,51200,2.088743209838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,3072,0.012191111014948951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,3072,0.01648000048266517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,3584,0.05599821938408745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,10240,0.0934248897764418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,64,0.017981333865059745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,65536,0.032427555984920926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,3584,0.5065431065029568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,256,0.007988444632954067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,512,0.021810667382346258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,12288,0.3927893373701308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,5120,0.009287111461162567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,8192,0.03691555394066705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,768,0.011448000040319232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,256,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,3072,0.016653334101041157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,7168,0.12325244479709202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,10240,0.03545511100027297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,256,0.0047404443224271136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,128,0.00555377784702513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,3584,0.011732444167137146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,2048,0.030713776747385662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,6144,1.7452418009440105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,5120,0.08059200313356188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,1024,0.008784888519181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,2048,0.14641333950890437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,3072,0.018383999665578205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,64,0.0026151111556424033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,5120,0.57587644788954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,6144,0.06087555487950643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,1024,0.01630577776167128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,3584,0.02606933315594991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,3584,0.06638489166895549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,3584,0.1074684460957845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,4096,0.07536888784832425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,12288,0.039621333281199135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,32,0.004609777695602841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,1536,0.035498665438758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,51200,2.713720957438151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,2560,0.07721510860655043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,51200,1.5055226220024958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,1024,0.013447999954223633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,768,0.017010667257838778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,8192,0.12987377908494738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,256,0.004392000122202767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,256,0.007156444092591603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,512,0.0058959999846087555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,2560,0.00962577760219574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,10240,0.696356455485026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,2048,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,3584,0.20781777964697945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,768,0.00831200016869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,3072,0.00815288888083564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,16384,0.1698231167263455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,3072,0.024097777075237695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,1536,0.025237333443429735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,64,0.0026062221990691293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,16384,0.2307057645585802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,3584,0.007897777689827813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,1536,0.004868444469239977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,512,0.006471111128727595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,12288,0.03717777795261807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,64,0.0041599998043643106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,8192,0.011081777513027191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,2560,0.012425777812798819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,128,0.009742221898502773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,4096,0.008778666456540426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,3584,0.22575110859341094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,3584,0.01499644418557485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,128,0.0052888890107472735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,5120,0.01276088919904497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,4096,0.45989688237508136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,51200,0.5868764453464085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,2048,0.011519110865063138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,3584,0.015477332803938123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,3584,2.8748914930555554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,3584,0.010926222635640038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,1536,0.04909422331386142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,32,0.008791999684439765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,3072,0.01716088917520311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,6144,0.019535111056433786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,256,0.04044533438152737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,256,0.004283555679851108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,128,0.01092355532778634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,6144,0.026465778549512226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,2048,0.283223999871148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,6144,0.024039111203617517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,256,0.02421333392461141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,16384,0.15590222676595053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,1024,0.009266667068004608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,51200,0.27103734016418457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,16384,0.133133331934611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,3584,0.10604800118340386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,5120,0.016859556237856548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,8192,0.006678222368160884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,51200,1.027168909708659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,12288,0.015576889117558798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,4096,0.01200177768866221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,51200,3.003265804714627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,2560,0.008658666577604082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,12288,0.8799440066019694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,8192,0.10368533266915216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,5120,0.009844444692134857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,16384,0.7398551305135092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,64,0.006025777922736273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,64,0.006090666684839461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,3072,0.009342222577995723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,6144,3.4130496978759766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,5120,0.39420445760091144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,3584,0.03295022249221802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,5120,0.1291031175189548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,768,0.008580444587601556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,32,0.007556444240940942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,1024,0.005425777700212266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,4096,0.008056888977686564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,16384,0.01902844508488973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,10240,0.027331555883089702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,128,0.007688889073001013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,2560,0.01406044430202908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,7168,0.12159289254082574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,51200,0.3904426627688938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,3584,0.6578044361538357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,2048,0.011095999843544431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,8192,0.11941333611806233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,1536,0.02219111058447096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,32,0.0032515554792351196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,4096,0.0494959983560774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,512,0.03349333339267307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,128,0.014140443669425117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,128,0.00282222218811512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,16384,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,65536,0.12298133638170029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,768,0.09299466345045303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,2048,0.007530666887760162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,5120,0.04295466674698723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,12288,0.16333156161838108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,4096,0.05253600080808004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,512,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,3072,0.5556889110141331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,12288,0.18506577279832628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,5120,0.13296977678934732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,3584,0.02048711147573259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,128,0.0068408888247278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,8192,0.016903999778959487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,16384,0.7272462315029569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,2048,0.04946310983763801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,7168,0.01925155520439148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,3584,0.01642933322323693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,2560,0.18823733594682482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,1024,0.004658666749795278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,51200,0.9831537670559354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,32,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,2560,0.0495671100086636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,768,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,256,0.13646222485436335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,16384,0.09003822008768718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,128,0.0028977776981062363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,2048,0.01962933275434706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,1024,0.01697688963678148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,5120,0.015637333194414776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,1536,0.0407955547173818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,3584,0.014800000521871777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,3072,0.05103466577000088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,128,0.0028168888141711554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,128,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,256,0.007087111473083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,2048,0.015304888288180033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,6144,3.2050594753689237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,1536,0.024442666106753882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,32,0.005059555586841372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,2560,0.007708444363541073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,256,0.0029440000653266907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,3584,0.0955431130197313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,12288,0.033334222104814314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,8192,0.6468035380045573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,32,0.003478222423129611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,3584,0.0984497798813714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,10240,0.6005848778618706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,768,0.007813333637184566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,768,0.07590755489137438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,4096,0.014843554960356818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,7168,1.6433404286702473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,51200,0.5103519757588705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,6144,0.054321779145134814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,5120,0.05370577838685778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,8192,4.510670979817708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,3584,0.05679111348258125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,768,0.004634666774008009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,3072,0.04320977793799507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,512,0.028276443481445312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,8192,0.05665600299835205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,16384,0.06520622306399874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,6144,0.018626666731304593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,7168,0.17394488387637666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,256,0.005533333453867171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,1024,0.006055111272467508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,1024,0.0040995557275083335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,6144,0.04706044329537285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,256,0.004705777598751916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,1536,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,768,0.003929777691761653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,16384,0.03972355524698893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,51200,22.335430569118923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,10240,0.022338666849666174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,64,0.0047884442740016514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,768,0.025017778078715008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,51200,0.07761422130796644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,16384,0.6455439991421169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,7168,0.1059911118613349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,3584,0.010623111493057676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,64,0.011390222443474663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,8192,0.05990133020612928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,128,0.0028435554769304064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,6144,0.00980711148844825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,7168,0.027831110689375136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,51200,0.06535377767350939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,8192,0.0203146669599745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,32,0.01588088936275906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,3072,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,1024,0.013815999858909182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,65536,0.01534755527973175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,256,0.0035235554807715942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,6144,0.010688888529936472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,12288,0.11678844028049046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,3072,0.014557333456145393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,512,0.004575110971927643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,2048,0.03899022274547153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,1024,0.004143110993835661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,4096,0.008207110895050896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,3584,0.015641777051819693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,128,0.008541333178679148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,2048,0.07548355393939547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,65536,0.6379040082295736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,6144,0.03193777799606323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,32,0.0030097777230872046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,16384,0.04689066608746847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,4096,1.0110836029052734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,512,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,51200,0.5670773188273112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,12288,0.32584799660576713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,16384,0.029260443316565618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,512,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,3072,0.010375111467308467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,3584,0.018908444378111098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,3584,0.011057777537239922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,1024,0.025655110677083332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,256,0.08594399690628052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,512,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,10240,0.014574221438831754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,7168,0.02753066685464647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,65536,0.06792889038721721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,51200,2.0855511559380426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,65536,0.048998223410712353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,512,0.011854222251309289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,10240,0.10392000277837117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,3584,0.015173332558737861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,51200,0.8399750921461316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,768,0.005940444353553984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,1024,0.009983110758993361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,65536,0.03616888986693488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,6144,0.015929778416951496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,5120,0.010458666417333813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,2560,0.007288888924651676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,8192,0.07492799891365899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,6144,0.028852442900339764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,512,0.005623111294375525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,256,0.007033777733643849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,2560,0.3671155505710178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,51200,0.04857422245873345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,10240,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,512,0.005783111270931032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,32,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,512,0.025202666719754536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,2560,0.15677689181433782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,6144,0.26929601033528644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,4096,0.020800888538360596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,12288,0.14132889111836752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,3584,0.01609511176745097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,65536,0.08983200126224095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,65536,0.05520800087187025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,5120,0.00664533343580034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,256,0.00574222207069397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,512,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,2048,0.01367200000418557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,512,0.023481777972645227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,64,0.0052968888647026485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,2560,0.007147555549939473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,32,0.041201776928371854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,10240,0.028262221150928076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,6144,0.030152890417310927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,512,0.13309244314829508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,3072,0.34568177329169375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,64,0.014652444256676568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,768,0.01890222231547038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,512,0.00963199966483646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,8192,0.07067200210359362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,1024,0.005468444277842839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,3584,0.007583111524581909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,1024,0.060321774747636586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,32,0.024487111303541396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,7168,0.010842666857772403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,6144,0.07572799921035767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,32,0.02272266646226247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,3584,0.019461333751678467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,1024,0.055389334758122764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,128,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,16384,0.059750219186147056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,16384,0.021307556165589228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,32,0.02482755482196808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,128,0.007888000044557784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,10240,0.04684622089068095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,8192,0.14772000577714708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,256,0.011012444065676795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,7168,0.018774222996499803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,3072,0.03288622366057502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,51200,24.176589965820312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,3584,0.012421333127551608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,2048,0.02387999991575877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,4096,0.016383111476898193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,65536,4.421351962619357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,12288,0.03184266553984748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,64,0.006818666640255187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,128,0.011607111328177981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,51200,0.11484355396694607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,12288,0.057665778530968555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,7168,0.4177991019354926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,128,0.00378933341966735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,768,0.023393778337372675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,8192,0.0209840006298489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,65536,0.8728248808119031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,8192,0.6386062304178873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,32,0.00461777796347936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,65536,1.7727759679158528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,4096,0.01847199930085076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,6144,0.021122665868865118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,2048,0.0068862222962909276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,6144,0.3956960042317708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,5120,1.031384891933865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,128,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,2048,0.007627555893527136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,512,0.0075342224703894714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,3072,0.04598755637804667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,51200,0.09431644280751546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,512,0.003478222423129611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,2048,0.0918204453256395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,8192,0.06561244196361966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,6144,0.18559022744496664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,128,0.004481777962711123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,2048,0.04103911254141066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,1024,0.07242044475343493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,512,0.01329066687160068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,8192,0.4003911018371582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,1536,0.053583112027910024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,16384,0.7931289143032498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,6144,0.1241297721862793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,128,0.0037253333462609183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,3584,0.2028026713265313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,10240,0.4642488691541884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,128,0.006621333460013072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,1536,0.005695999910434087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,12288,0.08127644326951769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,65536,0.8626542091369629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,256,0.004881777697139316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,1024,0.0074240002367231585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,128,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,2048,0.01221333361334271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,16384,0.1274648904800415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,3072,0.19649867216746011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,3584,1.6947884029812281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,5120,1.2394719653659396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,1536,0.007081777685218387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,51200,0.06719999843173556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,32,0.008108444511890411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,128,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,3584,0.07097155518001981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,12288,0.17282756169637045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,7168,0.15317866537306044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,4096,0.2287128766377767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,5120,0.13518933455149332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,2560,0.04142489035924276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,8192,0.01460622251033783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,1536,0.006123555617200003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,65536,0.3850453429751926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,65536,0.0668355557653639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,128,0.0027013334135214486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,3584,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,1024,0.026191110412279766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,256,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,512,0.007281777759393056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,7168,5.428981357150608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,65536,0.08369333214230007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,4096,0.010414222048388587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,256,0.15211021900177002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,3072,0.04962311188379923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,8192,0.03329688972897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,4096,0.010046222143703038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,2560,0.013459555804729462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,7168,0.017143110434214275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,2560,0.5546533266703287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,10240,0.05786133474773831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,32,0.007837333612971837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,2048,0.006979555719428592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,10240,0.6546613375345866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,128,0.0046844445168972015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,768,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,256,0.005617777920431561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,768,0.029895110262764826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,6144,0.016361777981122334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,2048,0.016748444901572335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,8192,0.008737777670224508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,12288,0.02375288969940609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,51200,3.8920495775010853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,1536,0.02348355617788103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,1024,0.00795022232664956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,5120,0.09794311390982734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,51200,0.41360089513990617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,3584,3.350565380520291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,256,0.015272888872358533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,4096,0.06893155309889051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,12288,0.14227022065056696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,2048,0.01958933307064904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,4096,0.03252355588807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,4096,0.6730862193637424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,64,0.015150222513410779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,65536,0.5083093113369411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,2048,0.1165155569712321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,2560,0.008535111116038429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,1024,0.00666933341158761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,256,0.0033253334048721525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,512,0.03554577628771464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,51200,45.15637546115451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,1536,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,65536,0.030202666918436687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,7168,0.016209777858522203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,512,0.07442755831612481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,3584,0.011699555648697747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,256,0.007993777592976889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,1536,0.1325191126929389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,8192,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,7168,0.07810399929682414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,7168,0.04411111275355021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,7168,0.08309066957897611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,65536,0.12158932950761582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,12288,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,256,0.02148977749877506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,2048,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,65536,1.2864231533474393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,12288,0.040692445304658674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,5120,0.01440000037352244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,2048,0.011002666420406766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,512,0.004898666507667965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,32,0.004578666554556953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,128,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,6144,0.09239821963840061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,512,0.0527724458111657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,5120,0.00925422211488088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,2048,0.006978666616810693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,3584,0.030099554194344416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,16384,0.2840595510270861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,1536,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,128,0.02052177819940779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,12288,0.006163555714819167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,12288,0.0240639994541804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,12288,0.04502044452561272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,16384,0.25531289312574595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,7168,0.06972533464431763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,4096,1.1508568657769096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,5120,0.0251893334918552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,32,0.006344888773229387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,32,0.012807110945383707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,10240,0.0385484430525038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,3584,0.0094373333785269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,768,0.011944000091817645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,3584,0.11141510804494222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,8192,0.14379378159840903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,32,0.007903111477692923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,16384,0.2710239887237549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,8192,0.07037422392103407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,5120,0.0454711119333903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,12288,0.13202577167087132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,8192,0.013887999786270989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,16384,0.6937866740756564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,32,0.007988444632954067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,7168,0.03977155685424805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,2560,0.016746666696336534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,512,0.003346666693687439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,12288,0.01514844430817498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,2048,0.018579555882347953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,3584,0.011229332950380115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,1024,0.007603555917739868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,16384,0.13381244076622859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,3072,0.09859644704394871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,768,0.038981331719292536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,6144,0.20751110712687174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,51200,1.868011474609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,4096,0.024148444334665935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,2560,0.09099644422531128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,12288,0.015581332974963717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,10240,0.03611377875010172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,64,0.018367111682891846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,7168,0.02416533397303687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,1024,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,6144,0.0848497748374939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,768,0.0037226668662495087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,32,0.006296000132958095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,5120,0.014446222119861178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,768,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,1024,0.007346666521496243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,5120,0.033979554971059166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,2560,0.02717422280046675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,5120,0.20721777280171713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,16384,0.030236442883809406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,2560,0.026782222919993933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,16384,0.04052622119585673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,128,0.002977777686383989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,4096,0.09232711129718357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,4096,0.01277955538696713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,1536,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,2560,2.3360489739312067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,64,0.003646222253640493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,16384,0.03972622089915805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,768,0.008455110920800103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,64,0.009480000370078618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,1536,0.05854666895336575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,1536,0.08265422450171576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,51200,0.10533155335320367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,3584,1.4522061877780492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,1024,0.00535733335547977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,256,0.004558222161398994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,5120,0.0158942225906584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,768,0.028112000889248315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,1024,0.006102222121424145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,1024,0.025021332833502028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,256,0.0114702218108707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,65536,0.4687609142727322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,256,0.00573333352804184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,5120,0.014249778456158109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,256,0.006728000111050076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,12288,0.09145066473219131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,5120,0.08074400160047743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,7168,0.019000000423855253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,7168,0.0151555554734336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,3584,0.009676444033781687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,2048,0.009364444348547194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,65536,0.2523804505666097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,256,0.003152888889114062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,8192,0.032673776149749756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,5120,0.0111191107167138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,3072,0.0184906671444575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,128,0.20483999782138398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,512,0.2672213448418511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,8192,0.012130666938092975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,1536,0.008049777812427944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,2048,0.02839822239345974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,1024,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,3072,0.05565155877007378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,6144,0.0975688894589742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,10240,0.03456622362136841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,12288,12.837552388509115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,5120,0.05900622076458401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,5120,0.29895022180345326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,5120,0.011229332950380115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,51200,0.09545689158969456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,10240,0.0853759977552626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,32,0.002887111157178879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,65536,0.2421804534064399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,32,0.005513777749405966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,7168,0.02863022353914049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,4096,1.2942276000976562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,256,0.005063999858167437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,3072,1.815572526719835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,512,0.01313066648112403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,1024,0.06016800138685438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,3072,0.012418666647540199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,8192,0.1257608864042494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,3072,0.01678222252262963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,256,0.004412444515360726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,16384,0.1053511102994283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,512,0.020883555213610332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,7168,0.3067999945746528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,2048,0.056899554199642606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,1024,0.0052968888647026485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,32,0.0026213334252436957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,768,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,1024,0.005365333209435145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,7168,0.047170665529039174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,768,0.0185199992524253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,2560,0.016919111212094624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,12288,0.3984791172875299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,65536,0.09278666973114014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,4096,0.011015111373530494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,12288,0.06635466549131605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,16384,0.025176000263955858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,768,0.0061039999127388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,10240,0.14546755949656168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,4096,0.006234666539563074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,10240,0.11449510521358913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,768,0.004978666702906291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,12288,0.1800924407111274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,2048,0.0058959999846087555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,3584,0.2958746751149495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,4096,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,128,0.003993777765168084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,6144,0.03163466519779629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,7168,0.03597955571280585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,12288,0.043613334496816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,2560,0.014721777704026965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,768,0.020215999748971727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,12288,0.16184266408284506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,2048,0.03498044278886583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,6144,0.010191111101044549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,32,0.005211555709441503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,2560,0.01086400035354826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,65536,4.295199076334636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,5120,0.012974222501118978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,768,0.01665244499842326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,2560,0.016598222984208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,65536,2.0923750135633683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,65536,0.1030622190899319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,768,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,2560,0.008454221818182204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,64,0.010017777482668558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,3072,0.045385777950286865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,5120,2.435767067803277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,128,0.01164177805185318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,256,0.010271111296282874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,8192,0.4457973374260797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,768,0.012905778156386482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,1024,0.005635555419656966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,512,0.011426666544543372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,5120,0.17501955562167695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,6144,0.0888026687833998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,2048,0.016914667354689706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,10240,0.016711110870043438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,32,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,64,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,7168,0.1333271132575141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,10240,0.09347555372450088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,64,0.017670222454600863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,1536,0.060172445244259305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,65536,0.3108782238430447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,32,0.0033280000918441345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,1536,0.008750222623348236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,8192,0.1427688863542345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,768,0.15177155865563288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,128,0.005631999837027655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,1024,0.012260444462299347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,12288,0.4224257734086778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,10240,0.29400799009535045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,768,0.02089866664674547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,64,0.003914666672547658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,6144,0.03934933410750495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,65536,1.6497750812106664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,2048,0.018263111511866253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,256,0.06831377744674683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,10240,0.01997066703107622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,256,0.010926222635640038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,3584,0.17532800303565133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,4096,0.10264444351196289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,8192,0.015116444892353482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,7168,0.5523662037319607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,51200,0.3786639902326796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,2560,0.026172444224357605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,3584,0.0166631109184689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,6144,0.014865777558750577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,2048,0.03113599949412876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,10240,2.8543217976888022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,32,0.022557333111763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,512,0.005008000052637524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,64,0.01866044435236189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,3072,0.01752000053723653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,6144,0.021697777840826247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,2560,0.01942222151491377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,12288,0.6943475405375162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,51200,8.09759521484375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,1024,0.0053671110007498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,768,0.0074879998962084455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,512,0.009854222337404886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,64,0.0033368888414568375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,7168,0.017251556118329365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,256,0.04097422295146518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,3584,0.01866933372285631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,1536,0.0074444446298811175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,32,0.0027866665687825945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,64,0.010625777973069085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,16384,0.07398044400744967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,10240,0.33436889118618435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,128,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,1024,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,1024,0.7339359919230143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,128,0.004765333400832282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,7168,0.11635377671983506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,2048,0.02772177755832672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,2560,0.03072711163096958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,64,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,3072,0.16076444254981145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,16384,0.643862247467041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,65536,0.1714471048778958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,4096,0.016490666402710807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,2048,0.006843555718660355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,3584,0.007685333490371704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,768,0.04010399844911363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,768,0.005722666780153911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,256,0.003007111036115222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,32,0.0026657777941889236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,1536,0.9449609120686849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,16384,0.10088088777330186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,3584,0.008847999903890822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,16384,0.14219021797180176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,32,0.055904891755845815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,4096,0.025666667355431452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,6144,0.02832533253563775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,2560,0.027291556199391682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,1024,0.004047111090686587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,5120,0.057714667585161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,10240,1.5708871417575414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,10240,6.006360795762803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,2048,0.01181244436237547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,256,0.009964444571071202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,10240,0.030402666992611353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,32,0.004781333522664176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,7168,0.03756622142261929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,6144,0.37437955538431805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,2048,0.012073777616024017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,10240,0.019349333312776353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,4096,0.007267555428875818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,4096,0.014009777042600842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,10240,0.028845333390765723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,512,0.0058337777025169795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,32,0.009119110802809397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,2560,0.00942133367061615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,51200,1.6882479985555012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,32,0.003112000102798144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,5120,0.03856977820396423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,4096,0.01958755486541324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,10240,0.08610755867428249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,4096,0.011591999895042844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,1536,0.012873777912722694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,128,0.019834667444229126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,256,0.028455999162462022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,4096,0.5357555813259548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,4096,0.06092622545030382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,3072,0.009879111415810054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,256,0.004452444612979889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,512,0.005486222190989389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,65536,1.3316284815470378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,3072,0.024872000018755596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,512,0.008915555146005418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,256,0.003991111285156674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,51200,0.1379635598924425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,2560,0.02539377742343479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,64,0.005323555734422471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,32,0.014650666051440768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,256,0.012673777838548025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,3584,0.18002577622731528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,1024,0.004453333301676644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,256,0.0033128888656695685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,1024,0.006251555350091722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,6144,0.03489511211713155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,256,0.010738666686746808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,65536,0.3112906614939372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,64,0.005076444397370021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,128,0.11178044478098552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,768,0.00959288908375634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,256,0.002986666642957263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,2560,0.007239999870459239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,3584,0.2812480131785075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,128,0.010647111468844943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,16384,0.05292000042067634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,6144,0.03787022166781955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,8192,0.01789866718981001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,16384,0.04134044382307265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,768,0.007292444507280986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,7168,0.6649129125807021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,16384,0.07421333260006376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,6144,0.07748711109161377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,512,0.05102044343948364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,1024,0.14234844843546549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,12288,0.029711110724343195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,65536,2.1743492550320096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,256,0.006769777586062749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,1024,0.009152889251708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,3072,0.012650666965378655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,4096,0.03147466646300422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,2048,0.015218666858143277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,7168,0.011095999843544431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,128,0.008087111016114553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,5120,0.01421333352724711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,7168,0.012904889053768583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,6144,0.013442666994200813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,256,0.024187554915746052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,3584,0.1171591149436103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,1024,0.004892444445027245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,10240,0.014243554737832812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,10240,0.021737777524524268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,10240,0.24611377716064453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,2048,0.2141875558429294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,3072,0.821756468878852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,64,0.019296000401178997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,128,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,10240,1.8729805416531031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,1536,0.015899555550681222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,51200,0.075564444065094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,128,0.01904266741540697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,1536,0.01475644442770216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,6144,0.08746755785412258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,32,0.02781244450145298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,51200,0.5700311130947536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,2048,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,3584,0.041364444626702204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,51200,2.080679151746962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,16384,0.42542399300469297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,512,0.0090524446633127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,8192,0.037082665496402316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,2560,0.009764444496896531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,128,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,768,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,8192,0.01664177742269304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,6144,0.02679288884003957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,3072,0.061999109056260854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,7168,0.07228800323274401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,128,0.0041235557032956015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,256,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,1024,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,1024,0.00646577775478363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,4096,0.020662221643659804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,64,0.013306666579511432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,65536,34.392798529730904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,3584,0.09897955258687337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,2560,0.012888888518015543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,12288,0.29686400625440806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,5120,0.10736888647079468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,65536,8.154296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,7168,0.03091999888420105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,6144,0.10706044567955865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,3072,0.027768888407283362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,7168,0.017024889588356018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,7168,0.01496088835928175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,1024,0.006102222121424145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,7168,0.3306728998819987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,512,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,2048,0.09326933489905463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,256,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,32,0.0026764445420768526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,51200,0.7564213540818957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,256,0.10826044612460667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,16384,0.304696003595988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,12288,0.14087377654181585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,1024,0.011003555523024665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,5120,0.016314667132165697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,256,0.012823999755912356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,1024,0.003952888978852166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,10240,0.00941066692272822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,32,0.002704000100493431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,64,0.01461955573823717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,51200,0.01292711082432005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,5120,0.14014132817586264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,16384,0.2505626678466797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,2560,0.021858667333920796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,2048,0.009827555881606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,3072,0.007058666812049017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,3584,0.039078222380744085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,4096,0.0063742221229606206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,16384,0.08866933319303726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,768,0.029076443778143987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,7168,0.10071021980709499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,2048,0.879353788163927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,768,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,4096,0.18107288413577607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,16384,0.1272088951534695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,12288,0.49176708857218426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,256,0.021623111433453027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,2560,0.015433778365453085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,512,0.007176000210973952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,8192,0.015495111544926962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,3584,0.1003191139962938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,128,0.0026888888743188647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,768,0.3844284481472439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,65536,6.178799947102864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,3072,0.01198933356338077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,1024,0.009112888740168678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,256,0.009662222531106737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,3584,0.21419910589853922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,512,0.0063999998900625445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,128,0.014461333552996317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,1024,0.019869332512219746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,3072,0.0074026667409473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,32,0.01130844404300054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,51200,15.330166286892362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,7168,0.019716444942686293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,128,0.008807111117574904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,128,0.002801777794957161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,65536,0.7033599747551812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,768,0.005704000178310607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,1024,0.006753777878151999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,2048,0.05016888843642341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,2560,0.056403557459513344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,128,0.007000889215204451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,4096,0.21726577811770967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,512,0.028062221076753404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,12288,0.08316177792019315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,65536,0.6562568876478407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,10240,0.024466666910383437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,512,0.005036444299750858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,3072,0.08879111210505168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,128,0.14142666922675237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,4096,0.014192000031471252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,4096,0.08994311094284058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,8192,0.2269902229309082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,7168,0.06523466772503324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,768,0.02193333374129401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,768,0.006694222076071634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,10240,0.011608888705571493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,12288,0.019521777828534443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,3072,0.29608623186747235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,16384,0.1927351156870524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,2048,0.011848889291286469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,8192,0.06321333514319526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,1024,0.008904889225959778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,1536,0.01688799924320645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,7168,0.016543111867374845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,4096,0.017533333765135873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,65536,0.4856559965345595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,2560,0.00609422226746877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,4096,0.008399110701349046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,256,0.005758222192525864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,5120,0.35907732115851504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,256,0.003629333443111844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,2560,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,1536,0.008236444658703273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,16384,0.029215110672844782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,7168,0.01679999960793389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,8192,0.029955555995305378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,512,0.0069351109365622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,2560,0.10926844676335652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,128,0.004499555461936527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,8192,0.01368799971209632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,512,0.0424124432934655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,65536,13.773549397786459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,12288,0.056120885743035205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,7168,0.007941332956155142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,128,0.003248888792263137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,65536,53.873138427734375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,4096,0.4788551330566406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,51200,1.5139226913452148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,12288,0.10633599758148193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,7168,0.032039112514919706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,3584,0.023847111397319373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,5120,0.1530639992819892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,768,0.0457911127143436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,1536,0.02129688858985901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,4096,0.0188719994492001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,3072,0.01977688901954227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,1536,0.6900089051988391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,10240,0.006617777877383762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,256,0.004693333473470476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,51200,0.13851732677883571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,256,0.006779555645253923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,1536,0.013196444345845116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,65536,0.7255448765224881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,256,0.018754666050275166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,32,0.11485600471496582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,6144,0.024672889047198828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,3072,0.022142221530278523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,51200,2.9045999315049915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,65536,1.10388978322347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,1536,0.006060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,512,0.004962666581074397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,128,0.005040888984998067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,16384,0.5781857702467177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,6144,0.04314844475852119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,256,2.3756550682915583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,51200,1.049712922837999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,256,0.01012711144155926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,1536,0.008896888958083259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,65536,0.10787466499540542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,7168,0.4663777881198459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,64,0.0028239999794297745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,1024,0.02946044339074029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,512,0.004650666481918759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,5120,0.006923555499977536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,51200,0.2891484366522895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,65536,0.8228764533996582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,10240,0.4527609083387587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,256,0.004534222185611725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,1536,0.01683911184469859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,128,0.038943999343448214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,16384,0.03217333224084642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,1024,0.005173333403136996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,7168,0.013200889031092325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,5120,0.009062222308582729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,3072,0.007430222299363878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,2048,0.007652444144090016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,16384,0.16858488983578154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,3584,0.10425777567757501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,2560,0.011782222323947482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,3072,0.006969778074158563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,128,0.005921777751710679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,32,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,7168,0.01146933353609509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,65536,32.11116197374132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,7168,3.0802826351589627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,2560,0.007833778030342525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,16384,0.023967999551031325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,256,0.003126222226354811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,5120,0.009321778184837764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,10240,0.34291911125183105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,10240,0.10767022106382583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,1536,0.012667555775907306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,7168,0.024600888291994732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,768,0.04041422075695462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,768,0.05247199866506788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,1024,0.3923368983798557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,512,0.04628800021277534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,64,0.005385777602593104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,1536,0.07454222440719604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,4096,0.009578666753239101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,2560,0.026683555708991155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,5120,0.9451564153035482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,5120,0.01019022199842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,6144,0.06949510839250353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,65536,4.008354610866971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,3072,0.01401511165830824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,3584,0.06778311067157321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,7168,0.655353758070204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,1536,0.0069226668112807805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,512,0.016196444630622864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,64,0.00267999991774559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,512,0.003330666571855545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,2048,0.02143200072977278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,32,0.10028711292478774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,1536,0.0063573333124319715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,2560,0.03241066800223456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,3584,0.017846221725145977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,3584,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,10240,0.20168799824184844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,5120,0.03160355488459269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,4096,0.026702221896913316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,10240,0.5155973434448242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,128,0.003240888938307762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,3584,0.03829511006673177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,6144,0.013948443863126965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,64,0.006797333144479328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,512,0.007162666983074612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,1536,0.011215111447705163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,2048,0.011758222348160215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,12288,0.006575111299753189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,51200,0.07055999835332234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,4096,0.011251555548773872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,5120,0.00997155573632982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,512,0.007439999944633908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,3584,0.00998044427898195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,512,0.005552000055710475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,6144,0.020077332854270935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,16384,0.5329190889994303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,128,0.00909511082702213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,3584,0.011391110718250275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,4096,0.13369154930114746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,1024,0.01204711116022534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,1024,0.008446222378147973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,768,0.012801777985360889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,2560,0.010151111417346531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,512,0.016381333271662395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,7168,0.028246220615175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,5120,0.011800000237094032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,8192,0.07714933156967163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,1536,0.0063902222447925145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,3072,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,3072,0.047008888589011304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,1024,0.02469066613250309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,2048,0.013638222383128272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,2560,0.015117333994971381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,1536,0.006648889018429651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,1024,0.009135111338562435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,4096,0.008064000142945183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,6144,0.010596444209416708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,2560,0.04217777649561564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,12288,0.19231200218200684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,128,0.00463733325401942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,6144,0.00814666681819492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,2560,0.011228444675604502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,5120,0.028522666957643297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,512,0.05505955550405714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,7168,0.053835554255379565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,10240,0.21371821562449136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,6144,0.44680531819661456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,2560,0.0413680010371738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,256,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,256,0.020599111914634705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,128,0.030392888519499037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,32,0.005397333453098933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,4096,0.018575110369258456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,512,0.01664622293578254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,8192,1.175136884053548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,2560,0.01937866707642873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,256,0.0302471121152242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,6144,0.025077333052953083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,1536,0.0058906666106647914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,2560,0.07309155331717597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,768,0.010599999792046018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,16384,0.2026240030924479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,65536,1.5992648866441515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,3584,0.061553776264190674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,6144,0.007183111376232571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,1536,0.008243555823961893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,1024,0.018586667047606576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,512,0.01564711166752709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,12288,0.023940443992614746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,1536,0.006302222195598815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,3072,0.08495199680328369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,7168,0.03621866636806064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,12288,1.9358861711290147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,6144,0.029161777761247423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,256,0.005224888937340842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,6144,0.028739555014504328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,4096,0.185080885887146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,5120,0.019799111617936026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,64,0.006439111298984951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,8192,0.03997955719629923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,512,0.005290666802061929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,8192,0.26161334249708385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,512,0.004483555754025777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,12288,0.018422222799725003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,32,0.002663111107216941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,128,0.0047120000753137796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,51200,0.06384888622495863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,32,0.06805777549743652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,2560,0.04100888967514038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,51200,0.18580622143215605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,128,0.0069155556460221606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,128,0.004482666651407878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,8192,0.03191377719243368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,6144,0.29200177722507054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,256,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,768,0.26930401060316295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,2048,0.7539520263671875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,5120,0.013211555778980255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,64,0.004202666795916027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,3072,0.05292710993025038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,512,0.012538666526476542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,256,0.006247111078765657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,768,0.006412444429265128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,16384,0.039841777748531766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,2560,0.009189333352777693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,3584,0.030088000827365454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,10240,0.11042044560114543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,32,0.014596444037225513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,64,0.005010666532648934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,5120,0.012422222230169507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,8192,2.9283243815104165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,51200,0.14759200149112278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,256,0.005053333524200651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,5120,0.022068444225523207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,10240,0.10640533102883233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,65536,0.6413502163357204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,2048,0.0069066666894488865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,64,0.015849777393870883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,1536,0.4291173352135553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,51200,0.2830275429619683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,10240,0.02050933407412635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,51200,1.4603155983818903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,256,0.005956444475385878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,2048,0.05775377485487196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,7168,0.011252444651391772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,8192,0.01521511044767168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,12288,0.020749333831999037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,3072,0.01180622229973475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,1536,0.012093333734406365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,51200,3.0242453681098094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,51200,0.343178669611613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,2048,0.01056977775361803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,8192,0.0299271113342709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,3584,0.007541333635648091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,12288,0.006893333461549547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,768,0.022128888302379187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,1536,0.006248000181383557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,3584,0.015111999379263984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,3584,0.28865689701504177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,10240,0.03952088952064514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,6144,0.013648000028398303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,7168,0.03780355718400743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,3584,0.18773955769009062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,1024,0.03201955556869507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,8192,2.43595208062066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,8192,5.3943909539116754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,3072,0.030026667647891577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,512,0.01054133309258355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,128,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,10240,0.1414008935292562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,128,0.005990222096443176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,32,0.0028524444335036804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,64,0.0266266663869222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,512,0.07148622141944037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,65536,0.07899555232789782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,4096,0.020634666085243225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,65536,0.2989342212677002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,8192,0.01386844449573093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,65536,0.06521599822574191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,256,0.020234665936893888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,1536,0.006154666758245892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,32,0.002576888849337896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,12288,0.6647573577033149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,6144,0.021583999196688335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,128,0.005126222140259213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,3072,0.01793955597612593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,128,0.003773333297835456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,32,0.01032977799574534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,3584,0.007900444169839224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,3072,0.053244445059034556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,5120,0.2513279914855957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,512,0.02217955556180742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,65536,0.756390200720893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,6144,0.06669333246019152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,8192,3.908821529812283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,2048,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,32,0.019966221517986722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,1536,0.03835377759403653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,4096,0.012711111042234631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,51200,0.05011999938223097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,64,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,12288,0.026887999640570745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,64,0.11712177594502766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,256,0.004020444634887907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,51200,20.862640380859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,16384,0.5279599825541178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,16384,0.09384000301361084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,128,0.006218666831652324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,6144,0.012322666744391123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,51200,2.965657764010959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,1536,0.015451555450757345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,7168,1.537953800625271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,4096,0.07674488756391737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,7168,0.016536888149049547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,768,0.004609777695602841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,32,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,3072,0.014178666803571912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,5120,0.5165831247965494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,5120,0.016170667277442086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,10240,0.0136835558546914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,1024,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,8192,0.08420444197124904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,3584,0.010078222387366826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,4096,0.023379556006855432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,4096,0.010766222245163388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,4096,0.07214311096403334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,3584,0.13774932755364314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,1536,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,10240,0.4400480058458116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,51200,1.487460454305013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,1024,0.037631998459498085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,128,0.006000888844331105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,6144,0.041167110204696655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,2560,0.011249778171380361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,1536,0.004310222135649787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,3584,0.02514222264289856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,512,0.014874666929244995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,768,0.009468444519572789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,7168,0.039432889885372586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,2560,0.1480248901579115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,64,0.007045333584149678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,3584,0.0893262227376302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,256,0.00416355538699362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,10240,0.13557067182328966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,2560,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,1536,0.005955555372767978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,2560,0.007142222589916653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,8192,0.4230782190958659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,768,0.018746667438083224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,768,0.00546577779783143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,768,0.005223999834722943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,2048,0.12836799356672499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,32,0.0029306666304667792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,5120,0.14606132772233751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,2048,0.020067556036843192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,1024,0.008199111455016667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,1024,0.0302471121152242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,768,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,7168,0.1372071107228597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,65536,0.1444915533065796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,4096,0.01741066740618812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,32,0.006694222076071634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,3584,0.09735200140211318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,10240,0.1518968873553806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,8192,0.03445422318246629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,16384,0.5262862311469184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,3584,0.2641795476277669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,64,0.013255111045307584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,32,0.005747555444637935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,5120,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,12288,0.059070222907596164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,16384,0.12366044521331787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,3584,0.09444799688127305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,3584,0.007915555602974361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,7168,0.01924000018172794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,3584,0.016380444169044495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,32,0.08847288952933417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,128,0.00565422202150027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,1536,0.012568888564904531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,3584,0.007978666987684038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,512,0.019147555033365887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,4096,0.007692444655630324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,10240,0.07838488949669732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,768,0.04494400156868828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,256,0.0031679999083280563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,32,0.22841066784328887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,16384,0.9631484349568685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,65536,23.669769287109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,2048,0.1509795586268107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,5120,0.10583021905687119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,3584,0.012768000364303589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,7168,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,65536,0.040010667509502836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,1024,0.010440000229411654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,128,0.0029484445436133277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,256,0.004107555581463708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,64,0.040179553959104754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,7168,0.027624888552559748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,2048,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,32,0.01551555593808492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,128,0.004367111043797599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,51200,0.13475110795762804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,4096,0.05848089191648695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,3072,0.021613332960340712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,1536,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,512,0.006109333286682765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,1536,0.017030222548378837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,1536,0.00909511082702213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,128,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,5120,0.047913776503668896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,768,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,6144,0.007712889048788283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,4096,0.019448889626397025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,51200,0.2304168807135688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,64,0.00675999994079272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,5120,0.07263644536336263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,10240,0.15092533164554173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,32,0.009132444030708736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,64,0.002765333279967308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,768,0.026027555267016094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,2048,0.021611554755104914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,64,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,1024,0.013222222526868185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,51200,0.05540711018774244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,768,0.009496000077989366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,3584,0.012247111234400006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,1536,0.004410666724046071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,1536,0.0614373352792528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,5120,0.047704888714684375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,512,0.04537333382500542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,2560,0.012350222302807702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,1536,0.02455377744303809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,2048,0.011490667031870948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,16384,0.08326755629645453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,8192,0.03576711151334975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,2048,0.017105778058369953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,7168,0.13035021887885198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,3584,0.016741333736313712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,32,0.012390221986505719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,7168,0.015579554769727917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,128,0.00453599997692638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,10240,0.049359109666612416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,768,0.057499554422166615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,16384,0.03798933492766486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,1536,0.006160000132189856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,16384,0.16059199968973795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,12288,0.04093777802255418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,768,0.007976888782448238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,2560,0.006084444622198741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,1024,0.0037208890749348533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,51200,0.21237777339087593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,32,0.002630222174856398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,64,0.007557333343558841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,3072,0.019444444113307528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,65536,0.06705422533882989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,1024,0.0147706667582194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,6144,0.01314133322901196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,32,0.005511999958091312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,8192,0.05797333187527127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,3072,0.010501333408885531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,3584,0.8605288929409451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,2048,0.024944000773959692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,7168,0.016128889388508268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,8192,0.030163556337356567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,512,0.011762667033407422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,51200,0.15415199597676596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,2560,0.009786666267448002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,128,0.00852177788813909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,5120,0.0127688886390792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,4096,0.031129777431488037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,6144,0.010004444254769219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,4096,0.012545777691735161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,6144,0.025063110722435847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,6144,0.5444204542371962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,51200,0.3581662178039551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,4096,0.36725510491265195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,512,0.005397333453098933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,7168,1.9300827450222438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,2560,0.011485333244005838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,512,0.01741422216097514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,1536,0.0069884442620807225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,6144,0.021094222863515217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,768,0.15305511156717935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,16384,0.07788888613382976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,2560,0.4941440158420139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,16384,0.03921688927544488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,512,0.005491555564933353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,2560,0.011214222345087262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,8192,0.8081946902804904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,8192,1.9643014272054036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,256,0.003903111235962974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,8192,0.029286222325430974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,10240,0.1966480016708374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,3584,0.06380266613430448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,1024,0.0324524442354838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,256,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,51200,0.059770663579305015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,2560,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,6144,0.166174226337009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,65536,0.14215911759270564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,4096,0.011924444801277585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,65536,9.140337626139322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,5120,0.095195558336046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,1024,0.02121688922246297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,8192,0.12097599771287705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,12288,0.006151999864313338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,32,0.008174222376611497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,512,0.0081377774477005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,12288,0.18407555421193442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,32,0.022237333986494277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,1536,0.9816693200005425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,16384,0.014879999889267815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,5120,0.05088533295525444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,10240,2.8155564202202688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,65536,0.0934186644024319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,1536,0.018890667292806838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,512,0.01016177816523446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,2048,0.37542933887905544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,2560,0.006512889017661412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,2048,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,65536,0.8052639961242676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,64,0.00902133352226681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,1024,0.0074311114019817775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,2048,0.011896000140243106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,1024,0.12035111586252849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,256,0.16249422232309976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,5120,0.015736889508035447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,16384,0.09568622377183701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,10240,0.22979111141628691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,1024,0.045127110348807446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,8192,0.043720000320010714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,3584,0.6943030887179904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,128,0.004615111069546806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,65536,0.11849599414401585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,1024,0.023698666029506262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,5120,0.013385777672131857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,1024,7.050805409749349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,5120,0.014504889647165934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,16384,0.3093591001298693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,10240,0.039846221605936684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,7168,0.14508267243703207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,768,0.013088000317414602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,10240,0.036249776681264244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,16384,10.79805162217882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,7168,0.042344888051350914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,64,0.008649778034951951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,3072,0.10754577981101142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,6144,0.017832888497246634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,12288,0.02667644454373254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,7168,0.016326222154829238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,8192,0.022621333599090576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,1024,0.004921777794758479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,1024,0.009102221992280748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,512,0.009728000395827824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,2560,0.11119022634294297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,8192,0.01991555591424306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,3584,0.019483556350072224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,7168,0.11436622672610813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,5120,0.009416888985368941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,5120,0.015841777125994366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,1536,0.027297778262032404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,51200,0.1333928902943929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,6144,0.07211733526653714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,12288,0.7335093286302355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,3072,0.4072231186760797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,16384,0.18314666218227812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,8192,0.03864977757136027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,512,0.006631111105283101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,7168,0.036723554134368896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,512,0.011968889170222811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,128,0.018059555027219985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,128,0.0028808888875775864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,4096,0.011771555576059552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,10240,0.013567999833159976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,6144,0.014653333359294467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,3072,0.013556444810496436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,3072,0.11102044582366943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,4096,0.04729955395062765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,16384,0.42421600553724503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,16384,0.7225173314412435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,10240,0.008472888833946651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,8192,0.023050665855407715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,5120,0.012498666842778524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,16384,0.0396053327454461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,512,0.0053546664615472155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,8192,0.08468621969223022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,1536,0.004693333473470476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,768,0.03410933415095011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,4096,0.008223111430803934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,51200,36.39002143012153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,32,0.009575111170609793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,6144,0.04996622271007962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,1024,0.0057591112951437635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,8192,0.017457778255144756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,32,0.002532444480392668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,5120,0.006937777830494775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,4096,0.01111822244193819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,16384,0.15101155969831678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,2048,0.1168560054567125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,16384,0.006133333262470033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,2048,0.1510782241821289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,128,0.004469333423508538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,4096,0.17668533325195312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,8192,0.38953600989447695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,3584,0.016390222642156813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,7168,0.020012444920010038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,65536,2.1257928212483725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,256,0.0058755555914507965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,7168,0.026625777284304302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,6144,0.37718134456210667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,8192,0.015769778026474845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,256,0.01089333328935835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,65536,0.067266669538286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,2048,0.03912622067663405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,32,0.007248000138335758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,51200,0.061680886480543345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,768,0.008393777741326226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,32,0.010195555786291758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,5120,0.008527999950779809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,1024,0.007908444437715743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,5120,0.056087997224595815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,128,0.0032960000551409195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,64,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,65536,0.09201066361533271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,768,0.0033742222521040174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,256,0.007884444461928474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,2560,0.015503111812803479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,16384,0.32133155398898655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,2560,0.018183999591403537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,256,0.00815288888083564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,1024,0.004049777984619141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,65536,2.2033803727891708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,5120,0.007745777567227681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,4096,0.01330488920211792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,8192,0.6380222108629015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,8192,0.018368000785509746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,768,0.01053066634469562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,7168,0.5563244289822048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,768,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,6144,5.068419562445746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,16384,0.2838044431474474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,512,0.005112888912359874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,4096,0.02027466727627648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,3072,0.010218666659461128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,512,0.006702222343948152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,3072,0.02840533190303379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,51200,1.4647422366672094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,12288,0.10435466633902656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,128,0.005502222312821283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,4096,0.1268399953842163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,3072,0.008049777812427944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,1536,0.01036711119943195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,3584,0.09923911094665527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,2560,0.7180559900071887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,8192,0.0314995547135671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,64,0.003007111036115222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,3584,0.059060447745853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,2560,1.3525226381089952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,6144,1.8652532365587022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,10240,0.07228177785873413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,65536,0.3098737663692898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,65536,0.06512977679570515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,4096,1.7371519936455622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,8192,0.059452447626325816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,2048,0.04582666688495212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,768,0.5377368927001953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,512,0.08684533172183567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,5120,0.060144000583224826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,2560,0.06312177578608195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,3072,0.00810222244924969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,256,0.05852178070280287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,10240,0.029236445824305218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,256,0.0034017778105205963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,64,0.006412444429265128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,10240,0.02087822225358751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,12288,0.23487022187974718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,10240,0.2822835445404053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,128,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,51200,0.10989689164691502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,64,0.016414221790101793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,64,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,2560,0.019621334142155118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,7168,0.008621333373917473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,2560,0.11421688397725423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,64,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,4096,0.1510551108254327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,4096,0.014347554908858405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,3584,0.016693333784739178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,1024,0.15663999981350368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,7168,0.1192951069937812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,6144,0.03482133481237624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,5120,0.06425066788991292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,12288,0.0321395562754737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,64,0.005099555684460534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,65536,0.06604711214701335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,256,0.026888888743188646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,64,0.0026488889836602737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,2048,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,51200,0.05514488948716057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,3072,0.006056889063782162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,2560,0.20227111710442436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,5120,0.016184000505341422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,12288,0.39168445269266766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,8192,2.3744542863633895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,2048,0.007314666277832455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,1024,0.00599644457300504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,5120,0.7815199957953559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,512,0.05987911091910469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,16384,0.7566844622294108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,2560,0.035848001639048256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,768,0.03318488929006789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,2048,0.0075253330998950535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,3584,0.018740443719757926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,3584,0.010573333336247338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,3072,0.00777511133088006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,128,0.03548977772394816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,7168,0.010106667048401302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,32,0.02349244389269087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,768,0.008743111458089616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,5120,0.12928533554077148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,10240,0.034244444635179304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,3072,0.15458222230275473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,7168,0.10363466872109307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,1536,0.06297689014010958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,5120,0.00906666699382994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,6144,0.14404443899790445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,12288,0.0728951096534729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,2048,0.09463644689983791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,3584,0.012464889221721225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,1536,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,32,0.0025840000145965153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,12288,0.02758755617671543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,512,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,51200,1.288858625623915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,12288,0.016255110502243042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,5120,0.19000444147321913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,7168,0.2365430990854899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,256,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,2560,0.013936888840463428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,2560,0.09333066807852851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,8192,0.05437155564626058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,256,0.004879111217127906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,6144,0.009925333162148794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,1024,0.008643555144468943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,6144,0.015423110789722867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,8192,0.09325066539976333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,65536,1.2542720370822482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,51200,2.6061253017849393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,12288,0.8820977740817599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,2048,0.028573334217071533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,128,0.0027297778675953546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,2560,0.008283555507659912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,6144,0.012832889126406776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,4096,0.020224889119466145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,2560,0.009027555584907532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,3072,0.1420782274670071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,768,0.013178666432698568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,32,0.037053333388434515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,4096,0.010247111320495605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,16384,0.14149688349829778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,51200,0.01479466590616438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,512,0.0033457777980301115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,1536,0.014541332920392355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,3072,0.016615110966894362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,6144,4.958569844563802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,10240,0.24942045741611055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,4096,0.01332444449265798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,768,0.03273066547181871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,51200,26.625144110785588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,8192,0.27889688809712726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,12288,0.07056444221072726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,128,0.16296799977620444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,2560,0.06013511286841499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,768,0.011571555501884885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,65536,9.954988267686632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,51200,0.06353244516584608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,51200,0.08699822425842285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,2560,0.014847111370828418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,1536,0.00792711145348019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,4096,0.008179555336634317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,51200,0.22988443904452852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,3584,0.00963644435008367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,2560,0.029934220843844946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,16384,0.018774222996499803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,6144,0.04083289040459527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,128,0.017356443736288283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,256,0.0028693332440323303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,2560,0.012072000238630505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,768,0.037132445308897234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,4096,0.013261333107948303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,64,0.006605333338181178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,12288,0.1819902261098226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,1536,0.0048071112897661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,128,0.027867555618286133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,65536,1.0001564025878906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,3072,0.01828622155719333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,16384,0.026345777842733595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,3072,0.03755199909210205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,51200,0.1900364425447252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,1024,0.12458577420976426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,16384,0.29341332117716473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,10240,0.045253333118226796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,1024,0.04010577665434943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,3584,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,16384,0.34107377794053817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,512,0.018013333280881245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,4096,0.007839110990365347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,5120,0.1678088903427124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,5120,0.05461333195368449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,3584,0.09404089053471883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,2048,0.008938666847017076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,16384,0.543400870429145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,768,0.004837333328194088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,12288,0.03735111157099406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,2560,0.0514008899529775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,2048,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,3584,0.007251555720965068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,12288,0.018153778380817838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,1024,0.013289777768982781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,3584,0.13579377863142225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,1536,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,10240,0.026224000586403742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,128,0.0047582222355736625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,768,0.2562755478752984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,8192,6.078899383544922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,128,0.0041315555572509766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,51200,0.10995911227332221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,3584,0.016360888878504436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,10240,0.24039644665188262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,512,0.006254222244024277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,51200,0.1992302205827501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,2048,0.01571999986966451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,64,0.002826666666401757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,2560,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,65536,1.5584275987413194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,32,0.0034151110384199354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,16384,0.3855742348564996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,3072,0.014309333430396186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,64,0.0870622197786967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,4096,0.12055022186703152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,3584,0.12519200642903647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,5120,0.03394488824738397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,10240,0.05217866765128242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,256,0.008018666671382057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,512,0.005680000202523337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,12288,0.13378577762179905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,16384,0.020495999190542433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,65536,0.20127377245161268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,256,0.09132799837324355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,3584,0.03012088934580485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,8192,0.009264888862768808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,32,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,128,0.003027555636233754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,1024,0.00886044485701455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,10240,0.3268817795647515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,5120,0.01239999963177575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,10240,0.32812711927625865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,768,0.0057848890622456866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,256,0.13729688856336805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,10240,0.011221333510345884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,4096,0.018763555420769584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,2560,0.01640977793269687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,768,0.003530666646030214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,3584,0.056112892097897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,512,0.003313777761326896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,3584,0.027496889233589172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,128,0.0995253324508667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,3072,0.08206222454706828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,512,0.005416000054942236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,2560,0.022348443667093914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,51200,0.0648017790582445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,256,0.01052266690466139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,768,0.004250666747490565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,8192,0.026102221674389307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,2560,0.14907999833424887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,65536,0.5775938034057617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,64,0.018227555685573153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,7168,0.056067552831437856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,1536,0.04982399940490723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,1024,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,512,0.03525866733656989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,5120,0.03131555517514547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,1024,0.004893333133724001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,3584,0.0656906697485182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,4096,0.10808089044358994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,3072,0.018399111098713346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,4096,0.11961955494350857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,256,0.0069395556218094295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,7168,0.026942221654786006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,5120,0.07740177710851033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,16384,0.02857066525353326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,3584,0.024413333998786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,4096,0.00629066675901413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,3072,0.9515262179904513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,1024,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,2048,0.12533511055840388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,3072,0.014328888720936246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,16384,0.022085333863894146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,8192,0.05773778094185723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,256,0.014945778581831189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,16384,0.028075555960337322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,1024,0.0099653336736891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,65536,0.9869172837999133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,12288,0.07673422495524089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,3072,0.008064000142945183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,1024,0.013075555364290873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,65536,0.123325334654914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,128,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,6144,0.01052177780204349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,1536,0.041096889310412936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,16384,0.018552889426549275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,6144,0.024297777149412367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,51200,5.404829237196181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,512,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,3072,0.025207999679777358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,6144,0.15228799978892008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,65536,0.8079511324564616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,2560,0.009679111341635386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,32,0.023038221730126276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,7168,0.04215111003981697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,6144,0.015357333752844067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,64,0.009097778134875828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,3584,0.008775999976529015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,512,0.026001777913835313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,16384,0.4929831292894151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,5120,0.13977332909901938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,8192,0.017624000708262127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,7168,0.27926845020718044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,7168,0.1457200050354004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,1536,0.2900426652696398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,3584,0.012536889149083031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,12288,0.01679022279050615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,7168,0.011081777513027191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,3072,0.016780444317393832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,6144,0.009399111072222391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,7168,0.023980443676312763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,768,0.015200000670221118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,1536,0.008022222254011366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,10240,0.03707289033465915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,768,0.01293777757220798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,12288,0.4618382453918457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,7168,0.3971564504835341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,65536,0.40559911727905273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,256,0.027816888358857896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,12288,0.05147555470466614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,2560,0.007668444679843054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,6144,2.173815197414822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,1536,0.006476444502671559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,10240,0.03558222121662564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,512,0.008004444340864817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,4096,0.020576889316240948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,2048,0.14290577835506862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,2048,0.03767199979888068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,2560,0.031416889694001936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,32,0.010964444114102257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,4096,0.010429333481523724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,10240,0.00620977787507905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,2048,0.006836444553401735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,1024,0.008423111504978603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,7168,0.07172444131639269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,65536,14.10945044623481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,2048,0.013819555441538492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,64,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,7168,0.021493333909246657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,2048,0.008025777836640676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,51200,0.06279466549555461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,8192,0.2777591016557482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,4096,0.03364355696572198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,64,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,65536,0.0765955580605401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,2560,0.028910222980711196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,512,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,7168,0.010792888700962067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,7168,0.12389866511027019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,6144,0.014664888381958008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,10240,0.058386663595835366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,16384,0.04869955446985033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,8192,0.023784889115227595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,3072,0.018602665927675035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,128,0.005197333378924264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,8192,0.05855466922124227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,1024,0.02149155570401086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,65536,0.14590577284495035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,512,0.015815999772813585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,16384,0.48291911019219297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,16384,0.022364444202846948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,7168,0.051035556528303355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,16384,0.03786311215824551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,5120,0.011325333681371478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,2560,0.3386826780107286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,1024,0.011520889070298938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,12288,2.5092266930474176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,128,0.013968000809351603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,32,0.0031973332580592898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,1024,0.0064088888466358185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,4096,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,12288,0.11606400542789036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,64,0.012892444100644855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,8192,0.006623999940024481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,1536,0.005576888720194499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,10240,0.02254399988386366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,10240,2.995113796657986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,8192,0.17960445086161295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,7168,0.10798222488827175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,768,0.021252445048756067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,10240,0.10560622480180527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,256,0.005126222140259213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,1024,0.013516444298956128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,16384,0.07597511344485812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,2560,0.04101066788037618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,8192,0.11829155021243626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,8192,0.14651644229888916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,10240,0.10790933503044976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,3072,0.007407111426194509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,51200,0.17077688376108804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,3072,0.009345778160625035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,7168,0.06533066431681316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,10240,0.1886195606655545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,1024,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,7168,0.012068444656001197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,1024,0.012031111452314587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,5120,0.015746666325463187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,65536,0.10753777954313491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,768,0.007516444557242923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,256,0.002938666691382726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,512,0.28725332683987087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,512,0.005241777747869492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,16384,0.020979555116759405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,4096,0.020006222857369315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,65536,2.5397351582845054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,12288,0.19914400577545166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,2048,0.07842310931947497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,16384,0.022821333673265245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,6144,0.48264090220133465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,3584,0.03434755404790243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,768,0.012877333495352002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,6144,0.013418667018413544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,3584,0.010783111055692038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,2560,0.00812533332241906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,128,0.008998221821255153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,6144,0.0170133329100079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,64,0.0028071111689011254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,7168,0.1579022275076972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,7168,3.502655029296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,512,0.00811822215716044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,32,0.018756444255510967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,16384,0.05277688966857063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,5120,0.20137243800693086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,2560,0.09231999847624038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,5120,0.09773777590857612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,7168,11.621475219726562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,768,0.07379377550548978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,512,0.00832711077398724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,65536,4.03987545437283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,2560,0.17463466856214735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,256,0.010862222148312463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,1536,0.019122666782803006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,16384,0.026512889398468867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,256,0.007480888730949826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,10240,0.018220444520314533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,3072,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,32,0.002704000100493431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,6144,0.00666933341158761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,2048,0.00833422193924586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,32,0.015152000718646579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,7168,1.8598791758219402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,10240,0.1244915591345893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,512,0.017024000485738117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,1024,0.02160088883505927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,2560,1.3711342281765406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,4096,0.02279111080699497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,32,0.003378666730390655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,65536,0.25534221861097545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,1024,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,1024,0.01682933337158627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,128,0.06548711326387194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,2560,0.007794666621420119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,6144,0.2814577685462104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,5120,0.034303999609417386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,4096,0.7243359883626302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,3584,0.009681777821646797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,5120,0.010855110983053843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,2048,0.03802311089303758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,256,0.004421333471934001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,768,0.006145777801672618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,6144,0.015437333120240105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,1024,0.007073777417341868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,65536,0.008810666700204214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,1536,0.013189333180586496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,1536,0.006603555546866522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,65536,0.41185688972473145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,51200,0.11989511383904351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,768,0.02125066684352027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,512,0.008073777788215214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,1024,0.010008888940016428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,1536,0.1487440003289117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,2048,0.007256000406212277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,12288,0.5635911093817817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,3072,0.18692355685763887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,7168,0.02815377712249756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,256,0.004289777742491828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,4096,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,3072,0.04472444454828898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,128,0.0029582221888833572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,1536,0.004611555486917496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,256,0.003140444349911478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,7168,0.03463733196258545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,32,0.0035217776894569397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,64,0.004530666602982415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,64,0.0047528888616296975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,256,0.005356444252861871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,16384,0.07972444428337945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,32,0.006488889041874144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,64,0.005475555443101459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,3072,0.06602311134338379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,5120,0.009465778039561378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,128,0.005168000029193031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,16384,0.016580444243219163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,8192,0.00814488861295912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,128,0.0027226667023367355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,1024,0.007832888927724626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,12288,0.012167111039161682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,32,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,6144,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,2560,0.007050666544172499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,1024,0.11015822490056355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,5120,0.015757333901193406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,2560,0.011220444407727985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,3584,0.014890667464998035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,128,0.009364444348547194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,2048,0.007318221860461765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,16384,0.03951377669970194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,7168,0.013582222163677216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,8192,0.08142577939563327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,768,0.01295733369059033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,8192,0.008833777573373582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,12288,0.14184178246392146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,3584,0.008637333081828224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,2048,0.024353777368863423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,1536,0.6653306749131944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,12288,0.812188466389974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,4096,0.013015111287434896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,3072,0.048208889034059316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,32,0.00479644454187817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,1024,0.007515555454625024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,10240,0.2801537778642443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,4096,0.024900444679790076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,2560,0.15842400656806097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,2560,0.010440888504187265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,65536,0.4964097870720758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,51200,0.1393351157506307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,4096,0.10209066338009304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,8192,0.018592000007629395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,51200,0.39783377117580837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,3072,0.1977155605951945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,32,0.0025617778301239014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,10240,0.02127200033929613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,768,0.008940444224410588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,1024,0.010016889207892949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,16384,0.22284089194403756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,512,0.005055111315515307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,128,0.013227555486891003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,5120,0.09031555387708877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,6144,0.01240888900227017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,51200,4.0320659213595915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,7168,0.036687112516827054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,1024,0.03564088874393039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,256,0.02664622167746226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,2560,0.031347556246651545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,6144,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,512,0.003395555540919304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,32,0.018985778093338013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,1536,0.009088888764381409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,8192,0.012088889049159156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,65536,1.2907004886203342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,7168,0.03094666533999973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,2048,0.016028443972269695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,2048,0.03342311249838935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,5120,0.03554133243030972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,256,0.004741333425045013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,256,0.005625777774386936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,512,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,32,0.006798222247097228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,10240,0.016770665844281513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,7168,0.006450666735569636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,7168,0.021464889248212177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,32,0.006792888873153263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,512,0.013330666555298699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,51200,4.106655968560113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,768,0.02174577779240078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,1536,0.35054487652248806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,7168,0.02702222267786662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,1536,0.013832888669437833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,256,0.005899555567238066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,2560,0.00832177781396442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,7168,0.029286222325430974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,10240,0.011396444506115384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,5120,0.024762666887707178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,6144,0.01681777834892273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,10240,0.05963999695248074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,2048,0.007416889071464539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,32,0.06327911218007405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,16384,0.07900800307591756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,5120,0.2468257745107015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,1024,0.01514488955338796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,8192,0.36943554878234863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,128,0.042056000894970365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,16384,0.45439555909898544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,4096,1.6987590789794922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,12288,0.11983999941084121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,768,0.00489155575633049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,16384,0.018395556343926322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,4096,1.2474844190809462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,768,0.009862222605281407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,7168,0.007672000262472365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,8192,0.03157688842879401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,1536,0.006034666465388403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,256,0.007484444313579136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,128,0.004097777936193678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,256,0.051426665650473706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,2048,0.0326551099618276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,256,0.017643555998802185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,2048,0.0485155549314287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,5120,0.14188533359103733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,12288,0.22375112109714082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,4096,0.01126577787929111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,5120,1.123262193467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,3584,0.8407208654615613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,6144,0.007314666277832455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,8192,0.24487998750474718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,1536,0.013454222016864352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,3072,0.0733226670159234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,8192,0.015867556134859722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,2048,0.021942221456103857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,256,0.010168000227875179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,128,0.002740444408522712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,1024,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,51200,1.8449581993950739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,12288,0.01753777762254079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,256,0.023649776975313824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,512,0.005041777673694822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,8192,0.025235555238193933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,12288,0.07728799846437243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,6144,0.01644444465637207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,2560,0.007176888485749562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,128,0.004819555415047539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,2560,0.1425200038486057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,6144,0.023072888453801472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,64,0.00282577777074443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,2560,0.057004445128970675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,1536,0.005697777701748743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,8192,0.4153902265760634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,12288,0.28902310795254177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,2560,0.08533866537941827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,8192,0.04697688751750522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,64,0.005583999885453119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,2048,0.034790221187803484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,7168,0.017994667092959087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,51200,1.2090462578667533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,3584,0.05100533366203308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,3072,0.017224889662530687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,8192,0.00833777752187517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,1536,0.012995555169052549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,10240,0.01776622235774994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,6144,0.05010933346218533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,51200,6.030264960394965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,1024,0.023751111494170293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,12288,0.023190221852726404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,2560,0.007287999822033777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,10240,0.00692622239391009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,2560,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,3072,0.04088888896836175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,10240,0.10170844528410171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,51200,0.21943199634552002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,51200,0.22154577573140463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,8192,0.03356000118785434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,256,0.006692444284756978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,1536,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,51200,2.0168959299723306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,7168,0.2806684441036648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,768,0.005111111121045219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,10240,0.025570667452282373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,32,0.004132444245947732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,32,0.014670222997665405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,1536,0.009905777871608734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,51200,3.769036610921224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,5120,0.06614755259619819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,256,0.03151111139191522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,8192,0.034853332572513156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,10240,0.025849777791235182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,16384,0.06516533427768283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,1536,0.005033777819739448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,7168,0.01889244384235806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,2048,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,65536,6.649231804741754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,4096,0.02568088968594869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,8192,0.04824533396297031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,4096,0.031911111540264554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,5120,0.13610043790605333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,768,0.007147555549939473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,128,0.004415110995372136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,1024,0.00425155543618732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,7168,0.09288266632292007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,7168,0.022873777482244704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,4096,0.13090756204393175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,128,0.004302222281694412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,16384,3.0812863243950734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,2560,0.06864266925387912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,7168,0.014610666367742749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,12288,0.08441955513424343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,4096,0.009110222260157267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,65536,0.5535573429531521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,16384,0.02474311159716712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,16384,0.02072622213098738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,1024,0.006519111080302133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,12288,0.04308088951640659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,256,0.010087110930018956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,7168,0.036342223485310875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,6144,0.036248889234330915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,32,0.0032293332947625052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,51200,18.245961507161457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,768,0.009542221824328104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,2560,0.0659688909848531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,1536,0.0225564440091451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,3584,0.010980444649855295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,3072,0.015068444940778943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,512,0.17884622679816353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,3072,0.030294222964180842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,2048,0.1003040009074741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,3072,0.04455911119778951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,3584,0.008291555775536431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,128,0.009864889085292816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,51200,0.833463986714681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,256,0.01742577718363868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,12288,0.02520711057715946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,5120,0.022829333941141765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,512,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,16384,0.1844400035010444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,12288,0.5122355355156792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,51200,0.08515822225146824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,64,0.004543111142185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,64,0.03803022371398078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,12288,0.019368888603316415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,2560,17.759289211697048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,2048,1.4703280131022136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,1536,0.007852444218264686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,7168,0.07281244463390775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,32,0.005899555567238066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,12288,0.007105777661005656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,8192,0.042118221521377563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,8192,0.5529982248942057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,1024,0.004912000149488449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,12288,0.06778399811850654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,16384,0.07830133040746053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,4096,0.29206665356953937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,3584,0.03211377726660834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,5120,0.0069306666652361555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,2560,0.010608889162540436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,1024,0.007804444266690149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,32,0.01688088807794783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,65536,0.22395110130310059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,256,0.010693333215183683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,65536,0.7068799866570367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,128,0.009840889109505547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,2048,0.0421013335386912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,10240,0.04558133416705661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,3584,0.016749334004190233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,32,0.004401777767472797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,7168,0.23030845324198404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,8192,0.011679111255539788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,2048,0.0074275558193524676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,256,0.013174222575293647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,6144,0.06877244181103177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,3072,0.009935110807418823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,2560,0.008454221818182204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,768,0.04107022285461426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,32,0.002838222309947014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,2560,0.07719288931952582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,32,0.007747555772463481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,3584,0.010969777901967367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,65536,1.2844648361206055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,3072,0.014755555325084262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,1024,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,3072,0.017729777428838942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,2560,0.21072710884941948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,32,0.011580444044537015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,512,0.004326222257481681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,512,0.005709333138333426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,128,0.01699466672208574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,7168,0.0521048903465271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,2560,0.006474666711356904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,2560,0.006491555521885554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,5120,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,128,0.01402311192618476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,16384,0.022359111242824133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,512,0.005456888841258154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,12288,0.015425778097576566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,7168,0.014440000057220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,3584,0.4667439990573459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,65536,0.18635643853081596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,1024,0.06478222211201985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,768,0.016554666890038382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,32,0.0034337778472238113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,8192,0.1556817822986179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,1536,0.33872000376383465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,1536,0.13518577151828343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,8192,0.05048266715473599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,64,0.013096888860066732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,6144,0.010777778095669217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,3072,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,7168,0.11505511071946885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,3584,0.02518133322397868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,1024,0.004822222308980094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,3584,0.5999751091003418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,3584,0.045383999745051064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,5120,0.01093155559566286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,12288,0.06131911277770996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,3072,0.019109333554903667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,16384,0.1372257735994127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,7168,0.041329777903027005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,16384,0.029652443197038438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,16384,1.604538705613878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,5120,0.056094222598605685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,5120,0.008962666822804345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,128,0.0034302222645945023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,64,0.0029688889367712867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,2560,0.4277244408925374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,1536,0.010079999764760336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,3072,0.011058666639857821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,1536,0.01757955551147461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,1024,0.05747822258207533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,6144,0.859849770863851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,51200,0.06861333052317302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,16384,0.10993511146969265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,4096,0.19672800434960258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,16384,7.089458465576172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,768,0.054413331879509814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,10240,0.8359511163499621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,16384,0.46604712804158527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,51200,0.05899555815590752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,10240,0.007083555890454187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,256,0.007805333369308048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,10240,0.15615377161237928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,6144,1.268114619784885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,128,0.21203111277686226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,2048,0.009404444032245213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,768,0.007105777661005656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,51200,0.19504977597130668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,2048,0.005869333528810077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,12288,1.850189420912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,1536,0.014334221680959066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,3584,0.014661333627170987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,64,0.017815111411942374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,8192,0.02945599953333537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,51200,0.2603120009104411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,32,0.0034835553831524322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,8192,0.021075555019908484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,32,0.0036684444381131064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,3072,0.010199999643696679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,32,0.0024622221373849446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,10240,0.04968800147374471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,7168,0.01496355566713545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,2048,0.007450666692521837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,16384,0.7674764527214898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,1024,0.06413777669270833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,3072,0.021716444028748408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,8192,0.07856622007158068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,8192,0.010976889067225985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,512,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,7168,0.0140346669488483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,2048,0.007231111327807109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,1024,0.06607911321851942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,10240,0.07559733258353339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,65536,0.838447994656033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,1024,0.019279999865425956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,2048,0.01329066687160068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,6144,0.01038044442733129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,65536,0.09635644488864475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,51200,0.19101866086324057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,1024,0.00444088876247406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,6144,0.2055955595440335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,32,0.0027217778066794076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,10240,0.07665599717034234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,512,0.005599110904667113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,3584,0.14553688632117376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,4096,0.02737599942419264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,64,0.004552000098758274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,51200,0.1670008897781372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,768,0.11439555221133763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,1536,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,65536,0.0768151084582011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,1536,0.10423288742701213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,768,0.03118577930662367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,32,0.018604444132910836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,5120,0.03269155489073859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,1536,0.014870221416155497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,64,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,51200,0.42969245380825466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,768,0.010255110760529837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,256,0.01868088874551985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,2048,0.007712889048788283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,4096,0.014833778142929077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,7168,0.011584888729784222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,32,0.013450666434235044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,5120,0.07251555389828153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,6144,0.02184355590078566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,2560,0.019641778535313077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,2048,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,256,0.015744000673294067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,65536,1.2910168965657551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,1024,0.005789333333571752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,10240,0.008465777668688033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,51200,0.059487111038631864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,8192,0.12405778302086724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,12288,0.016199999385409884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,65536,8.283188713921442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,65536,0.7593190934922961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,64,0.002826666666401757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,51200,0.44272979100545246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,128,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,51200,33.56561618381077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,3584,0.013772444592581855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,32,0.032050665881898664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,2048,0.012444444828563266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,1536,0.007066667079925537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,5120,0.03240622083346049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,2048,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,51200,0.06819288598166572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,2048,0.037181334363089666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,3584,0.020091555184788175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,6144,0.05729866690105862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,6144,0.02090577781200409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,16384,0.6108249028523763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,16384,0.8857848909166125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,1536,0.011644444531864591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,2560,0.01148622234662374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,1536,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,65536,0.14140000608232287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,51200,0.134387559360928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,2048,0.058873779243893094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,51200,0.153583116001553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,6144,0.031543999910354614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,768,0.17395110925038657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,6144,0.02422933280467987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,2048,0.007146666447321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,8192,0.26235999001397026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,3584,1.3455502192179363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,1024,0.007089777953094906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,1024,0.009653333160612319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,32,0.01146488885084788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,3072,0.06698044141133626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,6144,0.013960889644092984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,7168,0.010882666541470421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,6144,0.26762400733100045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,1536,0.0069208890199661255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,1536,0.029329776763916016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,3584,0.007528000407748752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,2560,0.03619022170702616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,8192,0.04274310999446445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,1024,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,32,0.0041413332025210066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,51200,0.8369262483384875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,1536,0.014183999763594734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,8192,0.9210400051540799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,12288,0.17828444639841715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,8192,0.11812444527943929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,6144,0.015595555305480957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,1024,0.012168889244397482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,8192,0.012485333614879184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,65536,0.13384711742401123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,1536,0.009806222385830348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,1024,0.006583999842405319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,2048,0.007375111182530721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,512,0.008085333638721043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,16384,0.023158222436904907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,16384,2.7331155141194663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,6144,0.19480533070034453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,65536,0.48456891377766925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,2560,0.12410577138264973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,128,0.01255022237698237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,7168,0.014283556077215405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,768,0.06324711110856798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,128,0.005777777896987067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,4096,0.0196426659822464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,6144,0.10524000061882867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,256,0.005936000082227919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,16384,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,51200,0.26179199748569065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,768,0.0038773334688610504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,256,0.004408888932731417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,2048,0.18250577979617646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,32,0.004118222329351637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,512,0.005088888936572605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,5120,0.05298044284184774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,12288,0.17297511630588105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,3584,2.486202663845486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,128,0.005029333134492238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,8192,0.016255110502243042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,512,0.02244088881545597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,6144,0.010612444745169746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,65536,3.365305794609917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,10240,0.1555075512992011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,12288,0.3346702257792155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,64,0.006004444426960415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,2048,0.09941689173380534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,3584,0.09757333331637913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,8192,0.011519999967681037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,128,0.013615111509958902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,1536,0.015837333268589444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,32,0.016924444172117446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,4096,0.2769644525316027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,51200,0.059311111768086754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,1024,0.004041777716742622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,512,0.0033768889390759995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,3584,0.007923555870850882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,2048,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,5120,0.42851999070909286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,256,0.0027342221389214196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,3072,0.04696089029312134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,7168,0.3394631014929877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,12288,0.5268141958448622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,4096,0.008955555657545725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,128,0.0026906666656335196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,256,0.020448888341585796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,256,0.005320000151793162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,256,0.00349866681628757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,256,0.004680888934267892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,10240,0.22450844446818033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,1024,0.1390435563193427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,6144,0.05095200075043572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,2048,0.020727111233605277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,5120,25.034408569335938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,128,0.04538844360245598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,1536,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,7168,0.37163199318779844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,2560,0.014467555615637036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,12288,0.019136889113320243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,128,0.003605333467324575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,12288,0.05935111310746935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,128,0.009697777529557547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,1536,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,51200,0.18645866711934408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,1024,0.07901066541671753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,8192,0.01851733360025618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,51200,0.060263997978634305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,10240,0.04983733428849114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,3584,0.08244799905353122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,512,0.008435555630260045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,12288,0.5311084323459202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,2048,0.009708444277445475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,512,0.018073777357737224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,65536,2.5228390163845487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,2560,0.04036088784535726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,2560,0.008357333640257517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,2048,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,12288,0.0225591113169988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,7168,0.02090488870938619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,51200,0.30160178078545463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,16384,0.02475111186504364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,512,0.0388373335202535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,4096,0.5193884637620714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,65536,0.07483377721574572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,7168,0.05195644497871399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,1536,0.006783111227883234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,10240,0.1469351053237915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,2560,0.049048887358771436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,3584,0.04197777642144097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,16384,0.06106577979193794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,51200,1.058312839931912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,2560,0.01054044481780794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,768,0.25227377149793834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,16384,0.022600889205932617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,3072,0.057146668434143066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,8192,0.030973331795798406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,51200,0.10510489013459946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,7168,0.013591999808947245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,51200,0.05858133236567179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,16384,0.028889778587553237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,512,0.003415999934077263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,8192,0.030039999220106337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,6144,0.01148888882663515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,10240,0.043360888957977295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,16384,0.3672746552361383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,3584,0.012083555261294046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,768,0.36215911971198184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,512,0.013576889203654395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,128,0.006870222174459034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,6144,0.016892444756295946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,1536,0.019525332583321463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,1536,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,6144,0.011899555722872416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,768,0.006156444549560547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,12288,0.012295111185974546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,4096,0.03932444585694207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,64,0.012121777567598553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,5120,3.6599964565700955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,2048,0.025979555315441553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,128,0.0336444444126553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,128,0.15893956025441489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,768,0.009186666872766282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,1024,0.003974222060706881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,1536,0.03249333302179972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,16384,0.01016177816523446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,7168,0.08646400107277764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,2560,0.007170666423108842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,64,0.0036115555299652945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,4096,0.021799999806616042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,3072,0.0092613332801395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,51200,10.63087887234158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,10240,6.713725195990668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,768,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,5120,0.029776000314288672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,51200,0.06574577755398221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,256,0.008150222400824228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,2560,0.006992888947327931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,1024,0.007311999797821045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,10240,0.32838932673136395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,7168,0.15297510888841417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,51200,0.11571378178066677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,128,0.07873777548472087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,5120,0.0211208893193139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,5120,0.185151113404168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,12288,0.021772444248199463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,768,0.045977777904934354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,2048,0.006903111106819577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,3584,0.00888177752494812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,768,0.0035982223020659555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,12288,0.03128622306717767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,128,0.005531555662552516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,32,0.002740444408522712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,64,0.09465155336591934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,3584,0.008587555752860175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,2560,0.03049600124359131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,512,0.009117333425415887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,768,0.3288879924350315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,64,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,2048,0.17371555169423422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,128,0.015465777781274585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,10240,0.02664533257484436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,128,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,6144,0.010115555591053432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,8192,0.011785777906576792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,768,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,5120,0.3181564542982313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,7168,0.5332346492343479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,10240,1.3238755332099068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,3072,0.014249778456158109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,256,0.003907555507289039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,128,0.006151111175616582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,2560,4.312228308783637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,3584,0.18270488580067953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,128,0.0069191112286514705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,65536,11.212482028537325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,2048,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,4096,0.18147733476426867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,3584,0.07901155286365084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,4096,0.05970488654242622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,2560,0.17042133543226454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,128,0.057743999693128795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,768,0.20282310909695098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,512,0.02167555524243249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,512,0.0058115555180443665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,16384,0.04739200075467428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,2048,0.007640889121426477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,3072,0.021060445242457922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,3584,0.00812266684240765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,10240,0.06852622164620294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,128,0.07948622438642713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,5120,0.22255555788675943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,16384,0.021904889080259535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,256,0.006296000132958095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,16384,0.010496888723638324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,4096,0.03452800048722161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,1024,0.006713777780532837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,256,0.01496444476975335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,3584,0.23785156673855254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,12288,0.006999111009968652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,6144,0.006167999986145232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,1536,0.005707555760939916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,7168,0.031381332212024264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,1024,0.008616000413894653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,1024,0.005625777774386936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,6144,2.1728098127577042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,51200,1.035801781548394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,8192,0.20592710706922743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,3584,0.016178665889634024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,768,0.012713777522246042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,8192,0.07361421982447307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,4096,0.0476399991247389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,1024,0.004058666527271271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,6144,0.26145956251356334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,1024,0.0266951107316547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,7168,0.007287110719415877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,64,0.00546488869521353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,6144,0.030525333351559106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,51200,0.6149875322977701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,7168,0.40216888321770566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,65536,0.09793777598275079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,7168,0.19713866710662842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,1024,0.07724266582065158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,65536,0.8268239763047961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,768,0.0070364442136552595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,10240,0.07444710863961114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,2560,0.10536089208390977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,2560,0.05611911084916857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,5120,0.40216533342997235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,2048,0.024784889486100938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,768,0.021393777595625982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,65536,1.616970698038737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,1536,0.6168230904473199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,1024,0.02528088953759935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,10240,0.17770044008890787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,1024,0.05182844400405884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,1536,0.11061066389083862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,12288,0.14267822106679282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,12288,0.03750400079621209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,7168,0.4184586736891005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,8192,0.11514666345384385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,1024,0.02940800123744541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,8192,0.03512710995144314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,7168,0.07096266746520996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,5120,0.0185208883550432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,1024,0.04184444414244758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,32,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,4096,0.06122844749026828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,3584,1.298317379421658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,32,0.017216889394654166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,2560,0.011630222201347351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,5120,0.10087022516462539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,2048,0.01752711170249515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,1024,0.056009776062435575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,768,0.021038222644064162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,12288,1.914398193359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,3584,0.26523200670878094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,5120,0.015014222926563688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,51200,0.17632533444298637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,768,0.006413333531883027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,12288,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,1536,1.3847537570529516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,1536,0.025504888759719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,4096,0.14927732944488525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,32,0.0035173334181308746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,2048,0.054132444991005786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,32,0.008225777910815345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,768,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,256,0.008229333493444655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,6144,0.014153778553009033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,4096,0.014575999644067554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,2560,0.011519110865063138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,512,0.00870488915178511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,3584,0.00620266670982043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,10240,0.6702444288465711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,4096,0.32532533009847003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,4096,0.5194142129686143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,65536,0.20137066311306426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,8192,0.3421822124057346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,64,0.007198221981525421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,12288,0.014557333456145393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,512,1.9828915066189234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,8192,0.01677600045998891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,1024,0.021486222743988037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,6144,0.01614488826857673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,16384,0.878711117638482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,1536,0.03848000036345588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,65536,3.468975914849175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,2560,0.08471910821066962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,32,0.005009777843952179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,768,0.008772444393899705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,10240,0.020467556185192533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,32,0.06664711236953735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,1536,0.05632355478074816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,10240,0.13171288702223036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,16384,0.01570666664176517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,5120,0.08795911073684692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,12288,0.3264995680914985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,4096,0.011222222612963783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,3584,0.02606222199069129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,12288,0.04866400029924181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,256,0.003015999992688497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,512,0.011691555380821228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,512,0.014084445105658637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,2560,0.011132444772455426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,3584,0.010612444745169746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,7168,0.07851199971304999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,4096,0.5391964382595485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,12288,0.02310488952530755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,32,0.005988444305128521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,5120,0.009509333305888707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,65536,2.8944000667995877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,3072,0.013822221921549903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,2560,0.04529777831501431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,512,0.257238229115804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,1024,0.04376444551679823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,32,0.02555199960867564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,512,0.0035297779573334586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,12288,0.018069333500332303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,1024,0.003923555629120933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,512,0.03084533413251241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,64,0.002597333242495855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,2560,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,5120,0.01864622202184465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,512,0.0035573335157500375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,8192,0.12977688842349583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,16384,1.3214089075724285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,512,0.012663111090660095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,3072,0.06390400065316094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,128,0.004257777912749184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,3584,0.012658666405412885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,1024,0.014671110444598727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,768,0.0052044445441828836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,32,0.00350222239891688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,256,0.005500444521506627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,8192,0.12346577644348145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,768,0.052561776505576245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,256,0.2138542201783922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,6144,0.009541333549552495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,16384,0.025059555967648823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,3072,0.05571733580695259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,512,0.010825778047243753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,10240,0.11843378014034694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,8192,0.9631119834052192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,2048,0.03128711051411099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,51200,4.266320970323351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,10240,0.6124649047851562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,2560,0.014744889405038623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,1024,0.014071111877759298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,10240,0.024863999750879075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,2048,0.041512889994515315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,64,0.019348444210158456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,1024,0.02896089024013943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,32,0.0029395555870400537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,512,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,10240,0.030639999442630347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,8192,0.03972266780005561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,64,0.007091555330488417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,4096,0.11535733275943333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,512,0.013386666774749756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,512,0.28953777419196236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,6144,0.019807111885812547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,1536,0.009863110880057016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,65536,1.1636959711710613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,1536,0.00796088907453749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,7168,0.012051555845472546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,3584,0.07762400309244792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,5120,0.018741332822375827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,768,0.003563555578390757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,65536,0.7824017736646863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,32,0.029069334268569946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,16384,0.049379554059770375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,1536,0.007000889215204451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,5120,0.021028444170951843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,256,0.0037377778854635027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,51200,0.7829573419358996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,1024,0.015881778465376962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,64,0.008475555313958062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,6144,0.08486488792631362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,5120,0.014095111025704278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,128,0.004908444566859139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,7168,0.14756800068749323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,1024,0.010256888965765635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,256,0.008043555749787224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,3072,0.07783555322223239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,12288,0.02931555443339878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,16384,1.611544926961263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,768,0.0063591111037466265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,32,0.0033608888172441055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,128,0.005102222164471944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,1024,0.02206755512290531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,65536,0.05317688981691996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,5120,0.018018666240904067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,8192,0.019578667150603402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,3584,0.12368088298373753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,3584,0.008816888762844933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,6144,0.13777955373128256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,256,0.009175111022260454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,512,0.02564977771706051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,51200,6.457654317220052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,2048,0.08893244134055243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,3584,0.07514578104019165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,2048,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,2048,0.00628977765639623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,12288,0.059232890605926514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,256,0.005208000126812193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,1024,0.29768088128831655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,3584,0.07975021998087566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,1024,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,3072,0.010307555397351583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,5120,0.061536000834570996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,1536,0.0047528888616296975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,768,0.00565244464410676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,1024,0.04487466812133789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,8192,0.023203555080625746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,12288,0.04373688830269707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,1024,0.013417777915795645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,3072,0.006067555397748947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,1536,0.009768889182143742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,6144,0.06503911150826348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,3584,0.022237333986494277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,7168,0.029458665185504492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,7168,0.1153706709543864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,256,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,512,0.027318222655190363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,3072,0.03926488757133484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,8192,0.19678221808539498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,16384,0.0496053331428104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,8192,0.01418133411142561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,8192,0.03149066699875726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,32,0.005439110928111606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,256,0.010613333019945355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,2048,0.007728888756699032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,128,0.021389333738221064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,6144,1.5023991266886394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,6144,0.013470222552617392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,65536,10.576210021972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,256,0.009738666315873465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,8192,0.08695022265116374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,51200,0.3582453462812636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,512,0.006337777607970768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,512,0.008072888685597314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,16384,0.08999466896057129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,3072,0.011889778077602386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,32,0.011805333197116852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,128,0.007387555307812161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,10240,0.17823289500342476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,12288,0.01792355544037289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,65536,0.08940444389979045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,4096,0.5801982349819607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,3584,0.007903999752468532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,8192,0.054103109571668834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,5120,0.03194222185346816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,3584,0.025011556016074285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,768,0.023754666248957317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,32,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,32,0.03804355528619554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,8192,0.14519111315409342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,3584,0.10197510984208848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,51200,0.30273599094814724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,51200,0.10931466685401069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,128,0.004320000194840961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,2048,0.015973332855436537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,65536,0.14708977275424534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,12288,0.037053333388434515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,4096,0.014649778604507446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,32,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,51200,0.1790391074286567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,2560,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,16384,0.02012888921631707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,128,0.005094222310516569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,65536,3.9138514200846353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,7168,0.02158755560715993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,5120,0.008993777963850234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,10240,0.041794667641321816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,3584,0.014375999569892883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,2560,0.14543378353118896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,1536,0.009466666314336989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,7168,0.026133333643277485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,3584,0.018145778112941317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,512,0.010685332947307162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,5120,0.01921422282854716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,65536,2.400690714518229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,10240,0.02978666623433431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,2560,0.008913777768611908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,2048,0.016973333226309884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,3584,0.17344355583190918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,2048,0.009087111386987898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,10240,2.9535013834635415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,10240,0.3382151126861572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,6144,0.1689128875732422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,128,0.0058151111006736755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,6144,0.0739955570962694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,64,0.005677333308590784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,2560,0.006232000059551663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,12288,0.6223368644714355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,2560,0.05511555406782362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,5120,0.031705776850382485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,10240,0.14244622654385036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,8192,0.04320711228582594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,7168,0.021412443783548143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,1536,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,65536,0.16563110881381565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,3072,0.007932444413503012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,2560,0.009840000006887648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,768,0.004037333445416557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,2048,0.06910933388604058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,128,0.06376088990105523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,4096,0.009842666486899057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,16384,0.026446221603287592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,3584,1.8796043395996094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,5120,4.188163757324219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,1536,0.008422222402360704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,65536,0.15810133351220026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,65536,0.259007109536065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,3584,0.0184515549076928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,3072,0.008264000217119852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,1024,0.015352000792821249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,512,0.0092257774538464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,8192,0.30922754605611164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,256,0.007179555793603261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,2048,0.012999110751681857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,3584,0.3440222210354275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,768,0.024620445238219366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,2048,0.010620444185203977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,10240,0.02122488949033949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,256,0.00518844442235099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,1536,0.007931555310885111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,1024,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,4096,0.011576889289749993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,2048,0.012560000022252401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,32,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,1024,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,2048,0.004972444640265571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,256,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,256,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,4096,0.013183111117945777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,6144,0.3002222114139133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,51200,1.7557431326972113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,1024,0.0075431110130416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,32,0.008247999681366814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,8192,0.016876444220542908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,12288,0.12807022200690374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,8192,0.01186222251918581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,32,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,7168,0.010819555156760745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,10240,0.018736888964970905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,2048,0.043156445026397705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,16384,0.06615377797020806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,256,0.003032888803217146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,3584,0.01034755590889189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,2048,0.03630044394069248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,5120,0.9138995276557075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,32,0.00300177786913183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,768,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,3584,0.03535822365019057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,10240,0.02968977888425191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,1536,0.00498933345079422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,3584,0.00721688899728987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,768,0.004256000121434529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,256,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,1536,0.009280888570679558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,10240,0.03204800022972955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,2560,0.007667555577225155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,3584,0.007947555846638149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,4096,0.014916444818178812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,4096,0.019909333851602342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,7168,0.01089599976936976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,2048,0.016805332567956712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,65536,0.10043288601769342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,256,0.005523555394675996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,10240,0.5959324306911892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,7168,0.0923608872625563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,1536,0.008436444732877944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,768,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,2048,0.02068711154990726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,3072,0.014187556174066333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,256,0.003152888889114062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,12288,0.08093955781724718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,4096,0.3959395620557997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,6144,0.043508443567487926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,12288,0.037272890408833824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,5120,0.01310755560795466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,6144,0.008450667063395182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,10240,0.025588444537586633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,7168,0.1352942254808214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,3584,0.008873778084913889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,5120,0.07187288999557495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,8192,0.3340640068054199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,3072,0.6187662018669976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,1536,0.009405333134863112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,7168,0.014197332991494073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,64,0.002749333365095986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,2048,0.011744889120260874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,12288,0.07420355743832059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,51200,0.2793404526180691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,5120,0.013957333233621387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,3072,0.09392355548010932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,1024,0.01315111087428199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,768,0.006909333169460297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,768,0.006271111054552927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,10240,0.01920088960064782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,128,0.0032622222271230486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,1536,0.058851553334130176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,1536,0.030673778719372217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,1024,0.02109333376089732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,3584,0.14013511604732937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,5120,0.13388978110419378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,128,0.035509334670172796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,256,0.011368888947698804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,16384,0.025567111041810777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,5120,0.03820888863669501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,3072,0.011223999990357293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,64,0.002624888800912433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,51200,0.7201857566833496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,128,0.03673511081271701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,256,0.08190311325920953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,32,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,51200,6.964372422960069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,3584,0.020004444652133517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,512,0.024326221810446844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,1024,0.0088746663596895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,128,0.026515555050637987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,2048,0.03261422117551168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,2560,0.035098665290408664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,65536,0.07935288879606459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,128,0.4675591256883409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,4096,0.13115289476182726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,2048,0.010939555863539377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,1024,0.009321778184837764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,768,0.0036595554815398324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,51200,3.7278917100694446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,1536,0.016938666502634685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,16384,0.04962222112549675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,5120,0.06602222389645047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,6144,0.026978666583697002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,2560,0.00902844468752543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,1536,0.06751822100745307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,7168,0.04389066828621758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,3072,0.047873778475655444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,7168,0.014907555447684394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,7168,0.020011555817392137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,1536,0.00888533310757743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,10240,0.22587733798556855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,256,0.015358222855461968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,5120,1.7141235139634874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,16384,2.46741697523329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,2048,0.052523556682798594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,3072,0.06298755274878608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,512,0.01846844454606374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,512,0.011158222125636207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,2560,0.02269688910908169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,2560,0.025701334079106648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,65536,0.597832891676161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,768,0.009878222313192155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,51200,0.6383688714769151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,6144,0.03947644432385763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,5120,0.07465510898166232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,65536,0.10716444253921509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,1024,0.005678222411208683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,6144,0.028972443607118394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,12288,0.10403555631637573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,65536,1.7239831288655598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,51200,0.9812587102254232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,3584,0.009434666898515489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,32,0.0031964443624019623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,5120,0.02297244469324748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,8192,0.026872888207435608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,6144,0.0248364441924625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,65536,1.6336275736490886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,128,0.0063164445261160536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,5120,0.034061332543691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,3072,0.012568888564904531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,512,0.0035235554807715942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,4096,0.25288356675042045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,512,0.0558604465590583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,16384,0.020202666521072388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,2048,0.031785776217778526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,2560,0.007839110990365347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,12288,0.01590222285853492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,768,0.006406222366624408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,1536,0.01202311118443807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,128,0.003779555360476176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,3072,0.03216977914174398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,256,0.004624888714816836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,12288,0.05708444118499756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,10240,0.018159111340840656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,512,0.005776888794369168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,128,0.01878666712178124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,4096,0.019102222389645047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,65536,0.06867911418279012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,1536,0.009531555904282464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,1024,0.0053484443989064955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,51200,1.3213262557983398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,768,0.053800887531704374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,6144,0.026913776993751526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,8192,0.12661777602301702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,32,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,16384,0.06560711065928142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,5120,0.03766399953100417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,7168,0.31417867872450084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,3584,0.008251555263996124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,32,0.008565333154466417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,6144,0.05117866728040907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,8192,0.06195733282301161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,65536,0.07487733496559991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,1536,0.013600000076823764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,8192,0.43326489130655926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,6144,0.013556444810496436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,4096,0.007810666329330868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,512,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,51200,1.1738079918755424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,51200,0.19526667065090605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,7168,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,768,0.003590222034189436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,1024,0.021359110871950786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,2560,0.07836977640787761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,64,0.002708444371819496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,3072,0.007857778006129796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,8192,0.0617048872841729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,65536,0.9967750973171658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,512,0.01831644442346361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,1536,0.006224000205596288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,16384,0.02990488873587714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,256,0.005355555564165115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,3072,0.11650578180948894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,10240,0.009452444811662039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,6144,0.36391645007663304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,256,0.01641244524055057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,256,0.016083555089102853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,65536,0.09762844774458145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,10240,0.6869937578837076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,65536,0.009095999929640029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,128,0.0053839998112784485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,10240,0.08211821979946561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,16384,0.0692746639251709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,1024,0.012585777375433179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,1536,0.29352090093824595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,16384,3.4472293853759766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,3072,0.051860445075564914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,12288,0.034782220919926964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,5120,0.04252533449067009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,2048,0.011511999699804517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,8192,0.09130399756961399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,2048,0.04344355397754245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,51200,0.19884355862935385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,256,0.005848000033034219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,5120,0.0700942211680942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,4096,0.012472888661755456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,4096,0.012244444754388599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,10240,0.9245946672227648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,8192,0.11463110976748997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,5120,0.09090844127866958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,2048,0.008073777788215214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,768,0.005495111147562663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,2048,0.07419555717044406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,512,0.004498666773239772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,3072,0.03214666578504775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,3072,0.010300444232092964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,512,0.00600000015563435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,4096,0.12396177980634902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,3584,0.010877333581447601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,10240,0.043248888519075185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,10240,0.057144886917538114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,512,0.42319732242160374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,65536,1.9641910129123266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,10240,0.3192559878031413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,16384,0.034325334760877825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,12288,0.4181502130296495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,8192,0.013307555682129331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,256,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,3584,0.013481777575280933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,3072,0.015375110838148328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,4096,0.020440000626775954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,3072,0.05255377623769972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,2048,0.010681777364677854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,256,0.015708444847000968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,12288,0.4198026657104492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,768,0.07339821921454535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,3584,0.01959022217326694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,51200,0.10138844119177924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,5120,0.02942755487230089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,3584,0.052051554123560585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,65536,3.4008284674750437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,128,0.0030417777597904205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,768,0.0058648888435628675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,512,0.013648000028398303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,5120,0.4784364700317383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,64,0.0033315554675128725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,256,0.016007999579111736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,2560,0.018423999349276226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,16384,0.018585777944988675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,32,0.00740444411834081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,32,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,6144,0.019080888893869188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,1024,0.007223111059930589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,64,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,6144,0.0258577780591117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,7168,0.19505955113304985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,2048,0.07643289036220975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,51200,0.15777777300940618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,2560,0.01647822227742937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,6144,0.06858222352133857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,8192,0.40438932842678493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,64,0.020974222156736586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,2048,0.005415110952324337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,1024,0.009287111461162567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,1536,0.012301333248615265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,65536,0.36234132448832196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,7168,0.01182933317290412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,4096,0.008621333373917473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,2560,0.10017066531711155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,5120,0.11329955524868435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,12288,0.02884977724817064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,2560,0.009190222455395592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,256,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,5120,0.4667075475056966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,2048,0.013945778210957846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,51200,0.19828089078267416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,1536,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,10240,0.08587644497553508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,128,0.009147555463843876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,2048,0.1622942156261868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,1024,0.042414221498701304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,512,0.025775111383861963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,128,0.19756533039940727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,5120,0.026557332939571802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,2048,0.011676444775528379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,65536,0.24934845500522188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,2560,0.012028444144460889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,16384,13.961981879340279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,3584,0.013761777844693927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,12288,0.39515288670857746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,128,0.0033653332955307434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,1024,0.11083110835817124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,3072,0.022038223014937505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,128,0.0028417776856157514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,16384,0.04111822115050422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,5120,0.011796444654464722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,512,0.004171555654870139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,8192,0.007222221957312689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,6144,0.03308355477121141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,1024,0.36880355411105686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,16384,0.07591111130184598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,1536,0.019877332780096266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,64,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,1536,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,32,0.00526577772365676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,7168,0.12344443798065186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,5120,0.010313777459992303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,2048,0.5718586709764268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,512,0.03083555565940009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,2560,0.6893715328640408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,128,0.009016888837019602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,1024,0.046832889318466187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,2560,0.07360177569919162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,2048,0.01759466694460975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,2560,0.014420444766680399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,256,0.004822222308980094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,12288,0.02769333289729224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,256,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,10240,0.32965866724650067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,8192,0.02689244515366024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,5120,0.016189333465364244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,128,0.009903999666372934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,128,0.00517511119445165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,65536,0.09656266371409099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,3584,0.014896000425020853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,12288,0.07624977827072144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,3072,0.010247111320495605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,8192,0.05738844474156698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,3584,0.02625777820746104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,128,0.005285333428117964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,51200,0.09549599885940552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,5120,0.229747560289171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,1536,0.05519022213088142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,10240,0.031722668144438006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,2048,0.020503110355801053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,3584,0.022542221678627863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,4096,0.009463999834325578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,768,0.0034719999465677473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,2560,0.006160888820886612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,64,0.0030088888274298776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,1536,0.012006222373909421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,12288,0.14516266187032065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,4096,0.032834665642844305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,256,0.0073724447025193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,1536,0.019466666711701285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,7168,0.016898666818936665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,32,0.002862222285734283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,5120,0.012815111213260226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,64,0.005246222019195557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,2560,0.008765333228641087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,16384,0.1058880024486118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,128,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,5120,0.019723556107944913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,7168,0.34152621693081325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,3072,0.04236533244450887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,16384,0.07225777705510457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,12288,0.13648445076412624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,64,0.006678222368160884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,1024,0.004008000095685323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,1024,0.036355555057525635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,5120,0.03447555502255758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,12288,0.019816888703240287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,512,0.009611555271678502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,128,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,65536,0.07291111018922594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,51200,0.08072533210118611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,12288,7.230053371853298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,5120,0.04478044311205546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,1024,0.0317520002524058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,5120,1.6183137893676758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,64,0.003019555575317807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,64,0.005127999931573868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,32,0.0063582224150498705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,256,0.004287111262480418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,4096,0.03144444359673394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,1024,0.0047048889100551605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,64,0.005113777601056629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,8192,0.42154044575161403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,6144,0.08672355280982123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,512,0.30521154403686523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,768,0.01534844438234965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,768,0.006240889016124938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,6144,0.24395200941297743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,10240,0.010266666611035665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,4096,0.012627555264366997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,6144,0.05325422353214688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,3584,0.03030844529469808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,768,0.010635555618339114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,65536,1.731216854519314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,64,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,3072,0.012383999923865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,64,0.003363555504216088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,512,0.004523555437723796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,512,0.004835555536879433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,51200,0.25224444601270885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,12288,0.09436000055736965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,65536,5.746538798014323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,2048,0.008049777812427944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,128,0.009915555516878763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,256,0.02291999922858344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,768,0.14982755978902182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,51200,0.16859022776285806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,8192,2.629543940226237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,51200,0.08403111166424221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,1024,0.010748444332016839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,5120,0.10799644390741985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,3072,0.02662044432428148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,64,0.0030737777964936364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,65536,0.07506933477189806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,2560,0.015279110934999255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,12288,0.0347031123108334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,2048,0.00564177789621883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,1536,0.16521067089504665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,768,0.014139556222491793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,768,0.037400000625186496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,2048,0.06881244315041436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,32,0.007784888976150089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,16384,0.1827111111746894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,7168,0.015396444333924187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,2048,0.011913778053389655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,16384,0.032460444503360324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,16384,5.4234220716688375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,8192,0.03322044346067641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,32,0.011791110866599612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,128,0.006229333165619109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,128,0.004068444586462445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,512,0.006832888970772426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,10240,0.11232354905870225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,512,0.02195377813445197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,64,0.0075253330998950535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,3072,0.021087111698256597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,2560,0.6811084217495389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,3072,0.10715999868181016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,128,0.0038746665749284956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,64,0.017736888594097562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,16384,0.2323706679873996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,1536,0.03903822104136149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,4096,0.010876444478829702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,2048,0.00923466682434082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,8192,0.006599999964237213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,3584,0.0064479998416370815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,12288,3.0973750220404734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,2048,0.1496737798055013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,6144,0.34207823541429305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,65536,0.10253066486782497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,7168,0.023927110764715407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,1024,0.004299555387761858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,256,0.004500444564554426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,8192,2.0186773935953775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,51200,0.11371821827358669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,16384,0.031236443254682753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,51200,0.3120888868967692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,7168,0.02648977769745721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,2048,0.009992000129487779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,5120,0.014480888843536377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,3072,0.007547555698288812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,51200,0.054063111543655396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,64,0.0028195555011431375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,16384,0.07149155272377862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,6144,0.010858666565683154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,7168,0.019087110956509907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,7168,0.094687110847897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,3072,0.22820978694491914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,128,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,2560,0.11018755700853135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,12288,0.21221778127882215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,32,0.03863111138343811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,3072,0.008818666968080733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,256,0.04400177796681722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,10240,0.38893866539001465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,16384,0.049406223826938205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,512,0.015209777487648858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,2560,0.049955556790033974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,1024,0.035621331797705755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,768,0.0037226668662495087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,3072,0.01568800045384301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,51200,0.36043998930189347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,7168,0.016735111673672993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,2560,0.012067555553383298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,7168,0.2548533280690511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,1024,0.3170959949493408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,12288,0.0418524444103241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,128,0.00867377801073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,1024,0.058396445380316846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,128,0.003474666840500302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,16384,0.3729386594560411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,3584,0.008097777764002482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,2048,0.021881777379247878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,65536,0.6596417956882054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,16384,0.0288080010149214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,5120,0.04902844296561348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,51200,0.12111555205451117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,16384,0.009953777823183272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,256,0.004983110974232356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,10240,0.028332445356580947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,10240,0.019075555933846366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,16384,0.12390932771894668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,65536,0.11477066410912408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,768,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,1536,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,3072,0.031558222240871854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,8192,2.012178632948134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,256,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,256,0.017338666650984023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,65536,0.4482266638014052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,1536,0.006805333412355847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,256,0.006245333287451003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,3584,0.03912622067663405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,2560,0.01240355521440506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,4096,0.012532444463835822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,12288,0.10982222027248806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,32,0.009572444690598382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,7168,0.04490844408671061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,256,0.010096000300513374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,512,0.005626666463083691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,3584,0.23484444618225098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,6144,0.01108266661564509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,128,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,10240,0.042724443806542285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,3584,0.018917333748605516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,128,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,256,0.006665777828958299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,32,0.004181333300140169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,51200,7.690010070800781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,8192,0.018868444694413077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,65536,16.07184346516927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,51200,0.07984977960586548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,5120,0.025812443759706285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,64,0.016763556334707472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,32,0.054096887509028115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,5120,0.019494222270117868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,16384,0.00701333334048589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,32,0.00272444449365139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,7168,0.03135022189882066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,7168,0.10329155789481269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,3072,0.023744889431529578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,768,0.003560888684458203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,32,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,1536,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,3072,0.008236444658703273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,51200,0.10320533646477593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,64,0.002620444529586368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,16384,0.022991999983787537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,2560,0.23167999585469565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,1536,0.020491555333137512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,1024,0.005025777965784073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,12288,0.3181493282318115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,1536,0.006657777975002925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,128,0.007633777956167857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,3584,0.031005332867304485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,768,0.03194133440653483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,51200,0.053688890404171415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,65536,0.5480826695760092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,768,0.0196782218085395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,1024,0.005495111147562663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,256,0.005111111121045219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,8192,0.019751111666361492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,256,0.004214222232500712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,12288,0.06290933158662584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,4096,0.034874667723973594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,8192,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,1536,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,12288,0.18818667199876574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,256,0.006696888970004187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,768,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,65536,2.1780550214979386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,3584,0.04572355416085985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,6144,0.11842400497860378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,6144,0.07838222053315905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,128,0.004851555658711327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,6144,0.10862666368484497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,1536,0.0292559994591607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,3584,0.024488000406159297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,2560,0.03204888767666287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,256,0.011937778029176923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,6144,0.4128595458136664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,1024,0.008172444171375698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,2048,0.010686222049925061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,7168,0.00833333366447025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,16384,0.032123555739720665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,2560,0.008590222232871586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,4096,0.02680088910791609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,5120,0.022523555490705702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,16384,0.024975111087163288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,4096,0.012180444267061023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,1024,0.046613332298066884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,1536,0.004790222065316306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,768,0.004844444493452708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,3584,0.013951111170980664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,7168,0.015013333823945789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,51200,0.18447021643320718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,3584,0.0111191107167138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,6144,0.029450668228997126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,2048,0.006678222368160884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,65536,0.09160266982184516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,2560,0.00620888877246115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,1536,0.0047422221137417685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,768,0.06077510780758328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,1024,0.00609688874748018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,3584,0.012466666599114736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,2048,0.06022222174538506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,768,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,5120,0.10980711380640666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,2048,0.012415111064910889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,10240,0.5495173136393229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,6144,0.21180355548858643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,7168,0.02382755610677931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,3072,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,1024,0.006256000035338932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,51200,0.08183999856313069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,768,0.00979555563794242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,7168,0.13825155629052058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,3072,0.048435555564032666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,3072,0.015980444020695157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,65536,0.11275111304389106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,8192,0.2988995446099175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,5120,0.01017066670788659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,512,0.04196444484922621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,3072,0.03299644589424133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,768,0.003546666767862108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,3072,0.011312888728247749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,16384,0.09324977795283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,8192,0.07166666454739042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,256,0.004939555707905027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,8192,0.1491075621710883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,768,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,12288,0.019197333190176222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,128,0.0028728888266616394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,2048,0.024305777417288885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,4096,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,128,0.003160888950030009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,4096,0.014304000470373364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,6144,0.020738666256268818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,3072,0.03208266695340475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,12288,0.021596444977654353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,32,0.009291555318567488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,1024,0.009038222332795462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,512,0.039177778694364764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,2560,0.09768711196051703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,128,0.004489777816666497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,6144,0.03142133355140686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,512,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,1536,0.009133333133326637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,12288,0.018413333429230582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,3584,0.32422399520874023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,6144,0.02934755550490485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,1536,0.043488889932632446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,7168,0.031758222315046526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,5120,0.01330933305952284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,1536,0.03246488836076524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,65536,0.06270400020811293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,51200,18.045174492730034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,1536,0.007152888509962294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,1024,0.005443555613358815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,7168,0.0351493325498369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,8192,0.21511110994550917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,65536,0.0720720026228163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,16384,0.02102311121092902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,64,0.018820444742838543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,51200,0.41620800230238175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,64,0.03372800019052293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,2048,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,12288,0.008766222331258986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,4096,0.29489066865709096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,65536,0.13092888726128474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,16384,0.05023911264207628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,256,0.002926222152180142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,768,0.01331555595000585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,51200,13.097347683376737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,768,0.007807999849319458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,6144,0.05974666939841377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,3584,1.2361688613891602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,6144,0.04293955696953667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,2048,0.010065777434243096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,1536,0.029512001408471003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,128,0.0222080002228419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,12288,0.2830417686038547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,65536,5.4378416273329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,128,0.003269333392381668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,4096,0.05345066719584995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,2560,0.01278222186697854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,2560,0.4790533383687337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,7168,0.02016177773475647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,3584,0.054677334096696645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,51200,0.6110506587558323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,4096,1.9728480445014105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,3584,0.06216444571812948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,5120,0.01623911162217458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,3072,0.15018488301171198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,6144,0.36571110619439023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,128,0.004650666481918759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,16384,0.011709333293967776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,3072,0.011945777469211154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,65536,0.2060382233725654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,16384,0.42696889241536456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,768,0.006392000036107169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,3072,0.04158844550450643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,8192,1.2173129187689888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,3584,0.01310222182008955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,1536,0.01089866707722346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,16384,0.11448710494571263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,4096,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,7168,0.03856711255179511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,3072,0.015602666470739575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,12288,0.023413333627912734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,16384,1.1170061959160698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,1536,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,4096,0.01685688893000285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,64,0.01847466660870446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,8192,0.02053155501683553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,2560,0.015283554792404175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,64,0.00998222248421775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,10240,0.5490773518880209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,2048,0.0053119998839166425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,10240,6.838352203369141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,51200,0.11165777842203777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,64,0.006592888798978593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,1024,0.20527644952138266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,10240,0.026449778013759192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,51200,0.2074320051405165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,6144,0.08712444702784221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,1536,2.346184836493598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,64,0.0075253330998950535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,512,0.02011022302839491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,1536,0.10339466730753581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,768,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,1536,0.013071111506885953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,256,0.02517333295610216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,32,0.0037502224246660867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,512,0.006095110956165526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,4096,0.014143110977278816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,256,0.009704888694816166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,12288,0.053786665201187134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,51200,0.35038667254977757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,65536,4.162777794731988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,768,0.0038577777643998465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,6144,0.014597333139843412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,2048,0.1991964446173774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,7168,0.01890222231547038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,256,0.004915555732117759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,51200,0.362648884455363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,32,0.0026888888743188647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,3072,0.1331253316667345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,3584,0.04057066639264425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,3584,0.008024888734022776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,1024,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,6144,0.021357332666714985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,5120,0.012264000044928657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,2560,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,1024,0.0069919998447100324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,2560,0.008844444321261512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,2560,0.15610933303833008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,3072,0.10776355531480576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,4096,0.06769155793719821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,1536,0.005000888887378904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,3584,0.1228017807006836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,4096,0.030890666776233252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,7168,0.011946666571829053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,8192,0.013024888932704926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,8192,0.890334235297309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,51200,0.3834639920128717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,512,0.04584177666240268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,7168,0.030448890394634668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,1536,0.10435110992855495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,512,0.009175111022260454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,256,0.0032071111102898917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,128,0.0028213332924577924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,51200,0.7737626499599881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,5120,0.006804444309737947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,6144,0.013211555778980255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,12288,0.05377422107590569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,3584,0.02539644473128849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,768,0.006077333456940121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,1536,0.12009778287675645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,6144,0.11805421776241726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,2560,0.0074133334888352295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,1536,0.3352497683631049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,3584,0.1431866619322035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,1536,0.013286222186353473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,3072,0.05819466378953722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,2048,0.015049777097172208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,128,0.007306666837798224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,256,0.2316044436560737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,2048,0.10789510938856338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,5120,0.08452000220616658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,10240,0.02106311089462704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,2048,0.03477511141035292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,16384,0.008727110922336578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,3584,0.03054222133424547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,6144,0.16004355748494467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,5120,0.16517421934339735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,2048,0.015769778026474845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,7168,0.20407022370232475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,5120,0.041688889265060425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,1536,0.00461777796347936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,768,0.009703111317422655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,2048,0.01405066748460134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,16384,0.09090489149093628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,10240,0.06884533166885376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,512,0.0036862223512596558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,3072,0.06610400146908231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,6144,0.02363377809524536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,16384,0.09741244713465373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,5120,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,512,0.012170666621790992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,65536,0.07248800330691867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,3072,0.0625680022769504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,2048,0.027644443843099806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,1536,0.00981866651111179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,6144,0.028343111276626587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,128,0.004419555680619346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,1536,0.033434665865368314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,2048,0.005512889060709212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,16384,0.7179279857211642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,1536,0.015828443898095023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,2048,0.008791111409664154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,32,0.008719111482302347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,12288,0.019840000404251944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,32,0.003314666656984223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,7168,0.05355466736687554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,12288,0.016958221793174744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,3584,0.17094755172729492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,10240,0.03580266568395827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,512,0.027521777484152053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,51200,1.9108710818820531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,2048,0.008968888885445064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,12288,0.1426559951570299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,1536,0.0063235556913746735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,32,0.032553778754340276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,65536,0.25017423099941677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,6144,1.4005617565578883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,2048,0.012072888513406118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,65536,0.07877244551976521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,128,0.010298666854699453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,3584,0.01296533313062456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,1024,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,5120,1.0312897364298503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,768,0.021359999974568684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,12288,0.2853742175632053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,2048,0.16776000128851998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,51200,19.51180691189236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,16384,0.10061333576838176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,768,0.025892444782786902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,64,0.004184000194072723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,512,0.006204444501135085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,6144,0.015875554747051664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,7168,0.01091466678513421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,6144,0.0800959997706943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,65536,0.18437866369883218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,3072,0.007860444486141205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,51200,7.477108425564236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,1024,0.01406311160988278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,256,0.009854222337404886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,7168,0.034973333279291786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,4096,0.155149327384101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,2560,0.11332178115844727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,4096,0.028034667174021404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,65536,1.2266089121500652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,64,0.0030871110243929755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,4096,0.009555555880069733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,128,0.006872889068391588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,64,0.007473777565691207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,51200,0.33030666245354545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,5120,0.011676444775528379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,64,0.005411555369695027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,7168,0.1416248877843221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,128,0.002874666617976295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,3584,0.04466222061051262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,3072,0.00720088928937912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,4096,0.027971555789311726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,3584,0.06742578082614474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,51200,0.08332710795932345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,64,0.002983111060327954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,256,0.03510399990611606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,6144,0.012880889077981314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,512,0.0031946665710873077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,1536,0.011553777588738335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,1024,0.01814933286772834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,5120,0.4974657694498698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,7168,0.012455110748608908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,6144,0.14238311184777155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,768,0.011990221838156382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,7168,0.01463822192615933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,128,1.1829289330376518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,6144,1.8042320675320094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,128,0.0214373336897956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,16384,3.1203511555989585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,4096,0.02588533361752828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,6144,0.2058862182829115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,12288,0.20262932777404785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,128,0.019912888606389362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,10240,0.29084799024793834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,3584,0.9607413609822592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,6144,0.03817333446608649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,2560,0.024829333027203877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,4096,0.24283289909362793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,10240,0.019256000717480976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,2048,0.008087111016114553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,1024,0.007099555598364935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,768,0.20488177405463326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,2560,0.022621333599090576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,2560,0.01609599921438429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,256,0.005510222166776657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,7168,0.11897867255740696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,1024,0.006696888970004187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,65536,0.01811733345190684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,2048,0.11360266473558213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,1536,0.01111288865407308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,4096,0.03081866767671373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,768,0.03253511256641812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,768,0.003973333372010125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,8192,0.01572800013754103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,8192,0.022011554903454248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,5120,0.17798222435845268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,1024,0.036304887798097395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,2048,0.044713778628243335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,5120,0.47288089328342015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,3584,0.09535644451777141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,128,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,3584,0.08802666929033066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,2560,0.015963556038008798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,51200,0.0967902210023668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,6144,0.07123377588060167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,3072,0.1538506613837348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,65536,5.669737074110244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,128,0.009007111191749573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,1536,0.010281778044170803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,12288,0.021984888447655573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,2048,0.05355288916163974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,4096,0.006731555693679386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,16384,0.4253706667158339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,2560,0.14819822046491835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,768,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,64,0.04474311073621114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,256,0.015344889627562629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,512,0.01111822244193819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,1536,0.1079111099243164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,256,0.008391111261314815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,51200,0.07048533360163371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,8192,0.01703555550840166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,1536,0.0058240000572469496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,8192,0.03032711148262024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,128,0.006949333681000604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,10240,0.05351644423272875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,51200,0.03724977705213759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,8192,0.0159288893143336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,4096,0.04387822084956699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,1024,0.7897333568996854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,4096,0.06169155571195814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,128,0.006302222195598815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,1536,0.028808888461854722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,1536,0.06150488720999824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,128,0.00370755543311437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,128,0.003932444585694207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,32,0.010808889236715106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,65536,0.21138489246368408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,12288,0.03549066517088149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,8192,0.015005333556069268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,6144,1.1612648434109158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,768,0.010336000058386061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,16384,0.013003555436929068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,8192,0.2056666612625122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,3072,0.05290666553709242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,8192,2.1053377787272134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,2048,0.009729777773221334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,3072,0.028983109527164038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,128,0.0036106668412685394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,2560,0.01773955590195126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,5120,0.017400888933075797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,1024,0.04985688792334663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,16384,0.9598382314046224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,32,0.04925422204865349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,128,0.003439999909864532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,3584,0.03642222285270691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,1024,0.017118222183651395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,512,0.0034862222770849862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,16384,0.42148711946275497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,3072,0.033955554167429604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,16384,0.022618666291236877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,7168,0.013034666577974955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,6144,0.17090045081244576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,1024,0.013736888766288757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,6144,0.010120000276300643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,1536,0.012512889173295764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,128,0.012096000214417776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,512,0.005842666659090254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,1536,0.0277493331167433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,65536,0.06925421953201294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,7168,0.1692284478081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,1536,0.10225244363149007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,1024,0.010862222148312463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,256,0.0031102223114834893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,256,0.009705777797434065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,2560,0.1760257747438219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,16384,0.12252355946434869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,7168,1.349904908074273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,4096,0.11089066664377849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,1536,0.008261333737108443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,1536,0.01514933341079288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,8192,0.12033155229356553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,3072,0.03676355547375149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,5120,0.006609777609507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,1536,0.006127999888526068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,3584,0.007460444337791867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,1024,1.5017929077148438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,51200,0.0695306658744812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,8192,0.15359111626942953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,1024,0.012595555848545499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,2560,0.6496435801188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,2048,0.009880888793203566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,1024,2.7695164150661893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,65536,2.485990312364366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,128,0.019039111004935372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,2048,0.04179199867778354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,1024,0.04555288950602213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,1536,0.0075048887067370946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,2560,0.02603911028967963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,7168,1.2306924396091037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,7168,0.058543112542894155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,2048,0.23975910080803764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,512,0.01882133384545644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,10240,0.0210817770825492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,12288,0.038522667355007596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,3584,0.011899555722872416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,128,0.04807288779152764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,128,0.00923822240697013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,512,0.0034719999465677473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,2048,0.0324524442354838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,3072,0.006852444261312485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,1536,0.008800889054934183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,10240,0.03186310993300544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,3072,1.2720266977945964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,5120,0.052320887645085655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,65536,0.27662844128078884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,51200,0.21968089209662545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,2048,0.0074942219588491656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,6144,0.020311999652120803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,12288,0.010121777653694153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,1536,0.012506666282812754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,64,0.011155555645624796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,2560,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,51200,0.17011111312442354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,12288,3.7033776177300344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,7168,0.10901955763498943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,8192,0.026499556170569524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,64,0.004893333133724001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,65536,0.19167466958363852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,3072,0.015325332681337992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,768,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,51200,0.3183697859446208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,4096,0.009621333744790819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,1024,0.011585777832402123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,10240,0.3002071115705702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,512,0.00518222235971027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,4096,0.008981333010726506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,8192,0.08954222334755792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,4096,0.025274667474958632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,3584,0.0075635554061995606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,128,0.004560889055331548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,7168,0.587325308057997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,64,0.007785778078767989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,768,0.04123288724157546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,51200,0.5023075739542643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,6144,0.3651475641462538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,8192,0.27104711532592773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,5120,0.35321421093410915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,3072,0.0074471111098925276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,3072,0.014108444253603617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,7168,0.011139555937714048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,3584,0.014312000738249885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,16384,0.023191110955344305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,16384,0.4248666763305664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,12288,0.41829244295756024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,64,0.17480799886915419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,12288,0.2061448891957601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,128,0.005506666584147348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,128,0.0028533333291610083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,12288,0.5068008634779189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,10240,0.03242222136921353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,16384,0.02327466673321194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,64,0.009152889251708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,512,0.008257778154479133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,256,0.010117332968446944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,16384,0.05474844574928284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,5120,0.0803697771496243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,2048,0.04343733191490173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,12288,0.02016799979739719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,128,0.005993777679072485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,1024,0.033207999335394964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,5120,0.021328000558747187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,51200,0.6832222408718533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,2048,0.012987555729018318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,8192,2.2294364505343967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,1536,0.08503378099865383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,12288,0.048126220703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,512,0.003567999849716822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,16384,0.15080177783966064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,64,0.0027555556346972785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,4096,0.013967111706733704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,1024,0.005669333454635408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,128,0.00574222207069397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,5120,0.010468444062603844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,1536,0.0047973332305749255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,6144,0.024547555380397375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,32,0.017551110850440133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,2048,0.024783111280865137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,16384,0.7599875662061902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,1536,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,128,0.0034551111360390982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,2560,0.007385777930418651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,2048,0.04998666710323758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,2048,0.11928889486524795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,3584,0.023714666565259297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,256,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,3584,1.5956480238172743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,7168,0.35577599207560223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,6144,0.5174221992492676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,1024,0.020969778299331665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,12288,0.015751111838552687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,6144,0.10011821985244751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,2560,0.011200000014570026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,12288,0.036680890454186335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,6144,0.0466195543607076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,6144,0.011287111375066968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,16384,0.031692445278167725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,10240,2.576371510823568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,6144,0.043234666188557945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,8192,3.315063052707248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,16384,0.2833217779795329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,256,0.0041288890772395665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,256,0.018583110637134977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,7168,0.03769422239727444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,512,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,7168,0.23148266474405924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,65536,0.4866577784220378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,10240,0.03737777802679274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,8192,0.11873333983951145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,2048,0.02314311100376977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,65536,1.2689457999335396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,12288,0.02510133385658264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,64,0.006700444552633498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,6144,0.022156443860795762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,1024,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,8192,0.012207110722859701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,3072,0.015896888242827523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,12288,0.04612000121010674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,512,0.014318222800890604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,10240,0.3150640063815647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,512,0.02951999836497837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,1024,0.023356444305843774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,16384,0.31479644775390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,16384,0.022574222750133936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,512,0.016174222032229107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,12288,0.8049830860561795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,3584,0.02475555572244856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,2560,0.0844586690266927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,256,0.021669333179791767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,512,0.006908444480763541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,3072,0.0195448895295461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,256,0.03703555464744568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,128,0.004607999904288186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,5120,0.014316444595654806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,5120,0.02279288901223077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,7168,0.04862844281726413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,2560,0.014222222897741528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,8192,0.021780444516075984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,6144,0.011680888632933298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,51200,0.910180409749349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,256,0.008979555633332994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,1024,0.0075528886583116316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,768,0.06270311276117961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,5120,0.15324978033701578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,64,0.03638311227162679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,7168,0.011543999943468304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,4096,0.08262311087714301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,2560,0.037997331884172224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,65536,0.10806666480170356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,1024,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,7168,0.034916443957222834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,1536,0.015513777732849121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,256,0.010685332947307162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,65536,0.7757928636338977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,1024,0.015824889143308003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,1536,0.04943289028273689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,128,0.01203644441233741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,2048,0.00961599995692571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,768,0.005842666659090254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,1024,0.01089511149459415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,512,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,768,0.03616444600952996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,7168,0.018826666805479262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,8192,0.05177866750293308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,7168,0.02531822191344367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,32,0.0027093332674768236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,6144,2.1766622331407333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,4096,0.012859555582205454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,65536,0.08895111083984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,5120,0.012302222351233164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,1024,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,12288,0.05396355523003472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,10240,0.04199733336766561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,10240,0.055720885594685875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,7168,0.032348444064458214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,3072,0.011938666303952536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,16384,0.06991644700368245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,256,0.019128888845443726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,8192,5.359696706136067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,1536,0.00909511082702213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,1024,0.006873777757088344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,128,0.026344888740115698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,51200,0.04219466778967115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,8192,0.549365308549669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,2560,0.2152391142315335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,4096,0.009822222093741098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,10240,0.015256888336605497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,12288,0.7385555373297797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,7168,0.025455110602908667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,1536,0.018452444010310702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,2048,0.013624889155228933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,768,0.03705688979890611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,3072,0.008162666526105668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,8192,0.0768631100654602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,768,0.02179199953873952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,51200,0.08428000079260932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,5120,0.008754666480753157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,1024,0.010103999740547603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,12288,0.05363555418120491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,64,0.007704888780911763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,6144,0.11875643995073105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,64,0.00360977773865064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,12288,0.02590311070283254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,2560,0.005976888868543837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,8192,0.030863109562132094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,2048,0.007129777636792924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,3072,0.027237332529491846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,6144,0.023911111884646948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,10240,0.6790737575954862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,5120,0.020536889632542927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,2560,0.007659555309348636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,32,0.02977422210905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,6144,0.01772711177666982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,3072,0.011044444309340583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,4096,0.07449244128333198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,10240,0.3644728925493028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,4096,0.10590577787823147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,16384,0.018985778093338013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,51200,0.059234665499793164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,32,0.00675644435816341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,7168,0.020578665865792167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,3584,0.013198222551080914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,1536,0.0064080001579390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,2048,0.0464826656712426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,768,0.029969778325822618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,8192,0.03417511118782891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,3072,0.10955288675096299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,65536,0.137046217918396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,128,0.005599110904667113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,32,0.0041146667467223275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,6144,0.026389333936903212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,12288,0.03414933217896356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,65536,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,2560,0.010372444159454769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,65536,0.1287742190890842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,64,0.008170666793982187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,256,0.01068800025516086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,51200,17.39825608995226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,3072,0.04318311148219638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,128,0.005987555616431766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,1024,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,256,0.00300177786913183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,51200,0.04674577713012695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,4096,0.17143732971615264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,768,0.005418666534953647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,1536,0.0370506677362654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,7168,0.010967999696731567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,4096,0.02089244458410475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,512,0.00793688909875022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,8192,0.37749777899848086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,8192,0.44069600105285645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,12288,0.08322399854660034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,64,0.008169777691364288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,128,0.004493333399295807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,2560,0.04064177804523044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,32,0.04462933209207323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,128,0.02718488872051239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,16384,0.3301955593956841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,2560,0.009573332965373993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,16384,0.13839111063215467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,1024,0.00923466682434082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,128,0.013584888643688627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,32,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,512,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,16384,1.9019012451171875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,12288,0.05922489033805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,3584,0.025575111309687298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,1536,0.007983999947706858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,64,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,2560,0.019517333971129525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,10240,0.03791644506984287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,1024,0.01457333399189843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,2048,0.12156089146931966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,32,0.00598933340774642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,12288,0.037920888927247785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,10240,0.04909866717126635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,64,0.025124443901909724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,4096,0.23175199826558432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,8192,0.13714933395385742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,4096,0.04625422093603346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,1536,0.03563466668128967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,1536,0.0070053330726093715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,5120,0.1354124413596259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,2560,0.00796088907453749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,1024,0.006597333484225803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,5120,3.034371482001411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,2560,0.008893333375453949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,1024,0.012379555238617791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,3584,0.0965155561765035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,51200,0.8783377541436089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,65536,0.5452906820509169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,3072,0.022908444205919903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,2560,2.2172383202446833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,2560,2.1534656948513455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,10240,0.02124533388349745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,12288,0.04760711060629951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,7168,0.011552000211344825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,32,0.010080888867378235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,2048,0.04248266749911838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,512,0.06115111377504137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,16384,0.3862871064080133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,128,0.0040942223535643685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,2048,0.025993777645958796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,128,0.004274666723277834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,5120,0.015966221690177917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,2560,0.06953599717881945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,1536,0.006266666783226862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,128,0.09506577915615505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,256,0.003068444422549672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,65536,0.1226960023244222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,3584,0.017451556192504037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,16384,0.021020443903075323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,16384,0.02646222213904063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,8192,0.07976000176535712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,10240,0.017482666505707633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,6144,0.006241777704821692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,7168,0.008817777865462834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,3584,0.11649955643547906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,768,0.009383111364311641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,128,0.004479111068778568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,16384,0.19065954950120714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,4096,0.03924444317817688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,5120,0.009874666730562845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,6144,0.013620444469981723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,8192,0.023092443744341534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,12288,0.10338933600319757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,1024,0.1664702230029636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,12288,0.775138643052843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,10240,0.39413955476548934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,6144,0.12068711386786567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,12288,0.015086222026083203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,5120,0.03899911046028137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,256,0.0034088889757792154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,768,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,7168,0.022011554903454248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,3072,0.009504888620641496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,32,0.06769333283106486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,3072,0.020006222857369315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,10240,0.18296355671352812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,6144,0.01640266676743825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,128,0.02913155489497715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,6144,0.32569866710238987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,10240,0.057208889060550265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,16384,0.9847848680284289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,128,0.026878222823143005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,65536,0.13416177696651882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,6144,0.2749439875284831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,4096,0.008683555656009251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,3072,0.009961778091059791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,16384,0.06308355596330431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,5120,0.23005689515007866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,4096,0.08211821979946561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,5120,0.010590222146775989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,51200,0.1099475555949741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,1024,0.03241599930657281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,32,0.002623999905255106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,10240,0.012881778180599213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,2560,0.015156444576051501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,8192,0.0711688862906562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,1024,0.00793599999613232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,12288,0.018142221702469718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,3584,0.007502222226725684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,7168,0.09129955371220906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,65536,0.043050666650136314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,6144,0.1202239990234375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,2560,0.018838221828142803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,8192,4.971991221110026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,1536,0.009923555784755284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,3072,0.008805333740181392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,5120,0.2434257666269938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,4096,0.007932444413503012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,3072,0.0069288888739215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,5120,0.13485066095987955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,64,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,2048,0.014479999740918478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,2048,0.01899377836121453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,8192,0.020439111524158053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,3584,0.0064702220261096954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,10240,0.088919997215271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,1024,0.03568977779812283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,128,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,7168,0.037400000625186496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,3072,0.14697778224945068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,512,0.00847999999920527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,6144,0.35408798853556317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,3072,0.015127999915017022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,65536,0.4928160243564182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,12288,0.04257422354486254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,12288,0.016339555382728577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,128,0.006663110935025745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,512,0.011921777493423887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,64,0.1330097781287299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,16384,0.04244266615973579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,16384,0.37952621777852374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,10240,5.679533216688369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,128,0.009019555317031013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,65536,0.5032418039109972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,3072,0.056539555390675865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,32,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,3072,0.7018515798780652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,2048,0.009877333210574256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,16384,0.03299999899334378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,128,0.011754666765530905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,10240,0.02843466732237074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,256,0.004726222405831019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,16384,0.0381368895371755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,512,0.003628444340493944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,64,0.006228444476922353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,12288,0.22578133477105033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,768,0.003560888684458203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,128,0.00498577786816491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,4096,0.017662222186724346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,51200,18.706392076280384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,2560,0.016965332958433364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,256,0.0070773329999711775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,2560,0.008715555899673039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,256,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,512,0.03291555576854282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,7168,0.020728000336223178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,16384,0.06334400177001953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,5120,0.026404443714353774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,51200,5.161935170491536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,10240,0.1816462278366089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,4096,0.1534337732526991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,3584,0.15657777256435818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,512,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,1536,0.015379556351237826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,12288,0.10123822424146865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,3072,0.034953776333067156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,16384,1.2654462390475802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,16384,0.04133244355519613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,5120,0.013552000125249227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,2048,0.013843555417325763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,256,0.021381333470344543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,7168,0.6919137636820475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,16384,0.42493777804904515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,8192,0.014459555347760519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,4096,0.01977511081430647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,10240,0.16393066777123347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,16384,0.7360497580634223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,256,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,12288,0.04870311088032193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,6144,0.012582222620646158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,12288,0.13251288731892905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,2048,0.05989155504438612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,3072,0.007813333637184566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,6144,0.023962666591008503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,51200,0.12463200092315674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,2560,0.023170666562186346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,768,0.008868444297048781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,512,0.005471999860472149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,3584,0.010096888575288985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,6144,0.015311110350820752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,51200,0.09650311205122207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,64,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,6144,0.1532906691233317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,12288,0.925082630581326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,8192,0.15814844767252603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,1024,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,4096,0.011215111447705163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,1536,0.02204977803760105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,1536,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,16384,0.12310843997531468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,2048,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,6144,0.006501333167155583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,768,0.007884444461928474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,7168,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,3584,0.00850755555762185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,64,0.003077333379122946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,768,0.010816888676749336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,16384,0.014089778065681458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,7168,0.12152355247073704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,8192,0.17009688748253715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,5120,0.2800346745385064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,6144,0.15964355733659533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,1536,0.4592675632900662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,64,0.0026640000028742682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,3584,0.03249511122703552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,1536,0.02936533424589369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,4096,0.02370133333735996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,4096,0.027319111757808264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,768,0.007553777760929531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,6144,0.022904000348514978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,32,0.0028657778683635923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,3072,0.01369244439734353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,1024,0.005640000104904175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,2048,0.058118224143981934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,4096,0.008383110993438298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,768,0.005968889014588461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,8192,1.3119991090562608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,51200,0.06459289126925997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,2048,0.017045333981513977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,256,0.011906666888131036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,256,0.0035111109415690103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,7168,0.04840622345606486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,51200,0.29817599720425075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,128,0.00461777796347936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,512,1.829575008816189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,16384,0.02256977723704444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,65536,0.3518639935387506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,768,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,10240,0.027071999178992376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,1024,0.34639289644029403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,64,0.009134222235944536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,768,0.004001777619123459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,1536,0.02769600020514594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,8192,0.056254221333397754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,512,0.0069217777086628815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,12288,0.3712453312344021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,2048,0.034853332572513156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,128,0.006213333457708359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,3072,0.007368000017272101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,5120,0.006980444822046492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,3072,0.02640177806218465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,128,0.004350222233268949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,256,0.00555288874440723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,65536,0.009791110952695211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,3584,0.2061991161770291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,65536,8.024640401204428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,5120,0.13485066095987955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,128,0.02077244387732612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,256,0.005769777629110549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,16384,0.152547558148702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,768,0.033943110042148165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,65536,2.8877165052625866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,2048,0.12203466892242432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,3072,0.015132443772421943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,32,0.008284444610277811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,1536,0.07308266560236613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,65536,1.8918933868408203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,3072,0.04319733381271362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,7168,0.4660257763332791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,5120,0.13004799683888754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,768,0.02145333256986406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,3584,0.0107360002067354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,3584,0.07395466830995348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,16384,7.10916985405816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,4096,0.015304888288180033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,512,0.10045066807005142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,10240,0.04144000013669332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,4096,0.013728888498412238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,65536,0.534421337975396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,128,0.0033368888414568375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,65536,0.07516800032721625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,3584,0.01479111115137736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,1024,0.024666666984558105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,1024,0.022375111778577168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,2048,0.056925336519877114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,3072,0.01014311114947001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,1024,0.007087111473083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,12288,0.018391110830836825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,6144,0.02567822237809499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,5120,0.00924711094962226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,4096,0.1830408838060167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,512,0.025674665967623394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,768,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,7168,0.044124444325764976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,51200,24.074573092990452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,512,0.028973334365420874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,6144,0.220176895459493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,3072,0.19327910741170248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,128,0.005027555757098728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,512,0.003375111147761345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,51200,0.3079715569814046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,3584,2.1953866746690536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,1024,0.011114666859308878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,7168,0.02104533380932278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,8192,0.05894222524431017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,4096,0.009304000271691216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,256,0.007939555578761632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,65536,0.9973537656995984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,2048,0.008238222036096785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,6144,0.013304000099500021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,16384,0.06454755862553914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,2048,0.11257421970367432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,10240,3.199508455064562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,32,0.0033039999090962936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,12288,0.021363554729355708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,512,0.006897777732875612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,16384,0.5137146843804253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,512,0.006789333290523953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,6144,0.04681244492530823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,5120,0.1967386669582791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,3072,0.6038871341281468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,6144,2.3524275885687933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,51200,0.07682577768961589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,512,0.004519111166397731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,12288,0.01868266695075565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,65536,1.2258684370252821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,128,0.009710222482681274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,32,0.0033377777371141645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,2560,0.007735999921957652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,10240,0.0210826661851671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,6144,0.019333332777023315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,512,0.005790222022268508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,768,0.00906311141120063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,1024,0.010033778018421596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,65536,2.166510264078776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,3072,0.011819555527634092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,512,0.005427555491526921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,12288,0.715654214223226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,1536,0.01353422221210268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,51200,0.21074399683210585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,8192,0.12469155258602566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,10240,0.05033955640263028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,3584,0.011559999651379056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,3072,0.09428977966308594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,512,0.0031804444475306403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,51200,1.8288533952501085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,5120,0.08687644534640843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,7168,0.20857600371042886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,10240,0.010247111320495605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,2560,0.02871733241611057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,51200,0.2703813446892632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,256,0.0029235554652081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,3584,0.048785779211256236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,256,0.005512889060709212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,32,0.006719999843173557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,256,0.00889866633547677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,768,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,768,2.1206603580050998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,1536,0.056032001972198486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,3072,0.017394666870435078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,3072,0.02083555526203579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,512,0.010065777434243096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,1536,0.03278577658865187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,256,0.005416000054942236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,2560,0.035777777433395386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,12288,0.7319146792093912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,65536,0.2521431181165907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,2048,0.018027555611398485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,4096,0.021135111649831135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,51200,0.03103733393881056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,512,0.019837333096398246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,128,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,51200,1.2507173750135634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,12288,0.040969779094060264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,5120,0.029577778445349798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,16384,0.18168889151679146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,8192,0.07585866583718194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,5120,0.06343733602099948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,6144,0.031189332405726116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,512,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,3584,0.15013689464992946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,2560,0.06320622232225206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,1024,0.006179555422729916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,51200,0.05956177579032051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,1024,0.05633155504862467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,12288,4.387483384874132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,10240,0.13399377134111193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,1024,0.008709333009190029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,1536,0.004376888689067629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,2560,0.06020800272623698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,64,0.0026524443593290118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,4096,0.016221332881185744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,4096,0.2478560076819526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,256,0.01926577753490872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,3072,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,7168,0.2752506732940674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,4096,0.016559999850061204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,512,0.03991466760635376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,10240,0.029434667693244085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,7168,0.044738666878806226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,4096,0.013223999904261695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,512,0.021752888957659405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,10240,0.853507571750217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,6144,0.011894221935007306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,7168,0.07935288879606459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,2048,0.014875556031862894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,6144,0.029948443174362183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,12288,0.18742933538224962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,6144,0.05432977941301134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,65536,0.5053635703192817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,10240,0.14427910910712347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,128,0.06676710976494683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,65536,0.23342755105760363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,7168,0.010843555960390302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,1536,0.15994755427042642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,3072,0.021165332860416834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,7168,0.07447644737031725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,2560,0.02197422252760993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,1536,0.014107555150985718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,1536,0.020579554968410067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,1536,0.02241866621706221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,6144,0.007163555257850223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,256,0.0052844443255000645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,8192,0.029545777373843726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,256,0.10002222326066758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,4096,0.016506666938463848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,2048,0.015561777684423657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,32,0.007895111209816402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,2048,0.03997244437535604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,2560,0.011647111011876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,6144,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,512,0.0032346666687064697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,1536,0.18494310643937853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,6144,0.05221777823236254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,2048,0.016499555773205228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,2560,0.007558222446176741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,512,0.009402666654851701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,4096,0.1436577770445082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,16384,0.03598399957021078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,1536,0.0107360002067354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,65536,0.17290755112965903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,12288,4.0190959506564665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,2560,0.18509777386983237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,3072,0.01109688894616233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,2560,0.04142311215400696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,16384,0.0274000002278222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,12288,0.029811554484897192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,2560,0.03393244412210252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,51200,0.0863564411799113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,1536,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,7168,0.011034666664070554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,1024,0.05596622493531969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,2048,0.008777778181764815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,1536,0.01202222208182017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,16384,0.08928622139824761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,128,0.003958222352796131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,16384,0.018894222047593858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,128,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,51200,0.05842133363087972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,65536,0.35789865917629665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,1024,0.010096888575288985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,6144,0.035671111610200666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,8192,0.17114222049713135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,4096,0.0063946665161185795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,4096,0.10579911205503677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,64,0.016343999240133498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,768,0.02863022353914049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,2560,0.009559111462699043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,1536,0.009199110998047723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,1536,0.007239999870459239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,10240,0.2679671181572808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,768,0.008319999608728621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,2560,0.05304977628919813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,7168,0.18761600388420951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,1024,0.006244444598754247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,5120,0.12951733006371394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,128,0.0030257778449190986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,5120,0.31697156694200307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,3584,0.022392888863881428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,512,0.005922666854328579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,51200,4.03255123562283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,4096,0.03602755400869582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,2048,0.04262222184075249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,1024,0.010795555180973478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,7168,0.06366933054394193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,1536,0.01942488882276747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,512,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,6144,0.08214133315616183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,10240,0.029987557066811457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,2560,0.14002754953172472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,64,0.005335111171007156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,2048,0.006726222319735422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,7168,0.043859554661644824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,12288,0.015013333823945789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,10240,0.26460621092054576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,512,0.0047484445903036325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,2560,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,768,0.01218933363755544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,51200,0.5430151091681586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,2048,0.011306666665607028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,10240,0.2675057782067193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,1536,0.055388443999820285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,512,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,16384,7.146834479437934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,4096,0.7590071360270182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,3072,0.009173333644866943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,64,0.003142222141226133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,256,0.004783111313978831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,1024,0.005883555445406172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,8192,0.2999448776245117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,128,0.004053333153327306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,6144,0.013208000196350945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,10240,0.04311111238267687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,4096,0.15483289294772679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,51200,0.01864177816443973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,768,0.047651555803087026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,512,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,2048,0.008613333106040955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,32,0.014592889282438489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,64,0.002749333365095986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,512,0.01789066692193349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,12288,0.015203555425008139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,8192,0.03351377778583103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,2560,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,16384,0.015855999456511605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,2560,0.01718933383623759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,51200,0.015642666154437594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,8192,0.029542220963372126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,8192,0.13437333371904162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,1536,0.012182222472296821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,512,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,8192,0.00979288915793101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,4096,0.008801778157552084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,2560,0.25367644098069936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,3584,1.761219554477268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,5120,0.037243554989496865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,768,0.008551999926567078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,4096,0.013449778159459433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,2048,0.009381333159075843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,256,0.004423999951945411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,128,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,256,0.0031866667171319327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,5120,0.7260719935099283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,32,0.04784444305631849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,7168,0.3041280110677083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,12288,0.04342666599485609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,1024,0.010768888725174798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,2048,0.21860978338453504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,64,0.0027715555495686005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,1536,0.0064035554726918536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,1024,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,65536,0.09914666414260864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,2048,0.32597867647806805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,32,0.014525334040323893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,12288,0.029116445117526587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,64,0.002733333243264092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,8192,0.03456977672047085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,256,0.02350133326318529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,32,0.011920889218648275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,256,0.003112888791494899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,6144,0.03605955508020189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,7168,0.06970133384068807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,6144,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,2560,0.04666666520966423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,256,0.011818666425016193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,16384,0.022466666168636743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,2048,0.007717333734035492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,128,0.006679111056857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,10240,0.05973155630959404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,64,0.00702133360836241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,5120,0.07387822204165988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,65536,0.06951822174919976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,5120,0.2442399925655789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,51200,0.1810613340801663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,7168,0.10082222355736627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,10240,0.017089777522616915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,5120,0.006204444501135085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,6144,0.18652355670928955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,1024,0.01960000064637926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,3584,0.04205244448449877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,4096,0.10986133416493733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,512,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,8192,0.02773244513405694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,3072,0.09495377540588379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,128,0.004327110946178436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,3072,0.01184177812602785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,16384,0.04327199856440226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,128,0.005139555368158552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,32,0.0032595555401510666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,128,0.002732444347606765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,2560,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,5120,0.01183555523554484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,1536,0.04619377851486206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,12288,0.017445334129863314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,12288,0.027959111664030287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,3072,0.00813066628244188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,6144,0.10903822051154242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,12288,0.06881066825654772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,16384,0.4324471155802409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,1024,0.029348442951838177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,32,0.0024871110088295406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,3072,0.010820444259378644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,10240,0.01110755569405026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,2560,0.05264355407820808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,2048,0.06160089042451647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,16384,0.03825599948565165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,128,0.16800355911254883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,2560,0.375840875837538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,65536,0.22996621661716035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,10240,0.2471511099073622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,2560,0.04107644491725498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,256,0.00462666650613149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,32,0.010685332947307162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,2560,0.006728000111050076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,2560,0.6112862163119847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,1536,0.01257777793539895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,65536,0.33077777756585014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,2560,0.5403608745998806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,32,0.013791999883121915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,768,0.010712888505723743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,3584,0.7543759875827365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,51200,1.0669324662950304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,4096,0.009305777649084726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,1024,0.0064604443808396654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,12288,3.005213419596354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,2048,0.017282666431532968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,10240,0.027484445108307734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,5120,0.020761777957280476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,2560,0.04494844542609321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,10240,0.06487644380993314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,6144,0.30652443567911786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,3072,0.026945778065257605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,16384,1.230012469821506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,3584,0.01701955497264862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,1536,0.24965421358744302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,256,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,10240,0.025255110528733995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,8192,0.015803555647532146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,1536,0.10584088828828599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,256,0.005154666801293691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,8192,0.015075556106037565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,4096,0.010214221974213919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,512,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,2560,7.925631205240886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,512,0.011961778004964193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,6144,0.1236577828725179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,512,0.004825777891609403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,3072,0.009736888938479954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,5120,0.11651910675896539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,51200,0.054400000307295054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,3584,0.008621333373917473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,6144,0.009895111123720804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,3072,0.13539022869533962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,12288,0.017254221770498488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,8192,0.06257244613435534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,4096,0.04387999905480278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,1024,0.08771289057201809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,8192,0.4139511055416531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,8192,0.08437155352698432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,32,0.021998221675554912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,12288,0.06893333461549547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,2048,0.0687075588438246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,2560,0.1740666627883911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,16384,0.0225608895222346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,65536,0.07048444615470038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,51200,0.5548124313354492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,128,0.012061333490742577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,3072,0.2783786720699734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,256,0.0038035557501845886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,65536,7.119612799750434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,2560,0.8352568944295248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,10240,0.03958755400445726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,12288,0.01735822194152408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,51200,0.15716089142693415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,256,0.010679999987284342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,512,0.0033599999215867785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,64,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,10240,0.043876445955700345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,64,0.0046746668716271715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,3584,0.05194133188989428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,256,0.005515555540720622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,1024,0.020575111111005146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,51200,0.11362489064534505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,2048,0.012709333664841123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,512,0.009503111243247986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,3072,0.021128889587190416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,64,0.002959111084540685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,256,0.12067555056677924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,12288,0.31303466690911186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,5120,0.10908977852927314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,12288,0.03271111183696323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,1024,0.05292888813548618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,12288,0.2821235656738281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,8192,0.006712889091836081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,12288,0.07705510987175836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,10240,0.18681066566043428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,4096,0.13399377134111193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,3072,0.03187555405828688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,8192,0.4083973301781549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,6144,1.970950232611762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,3584,0.0325164430671268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,1024,0.010034666293197209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,5120,0.18489510483211943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,7168,0.17671288384331596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,1024,0.025210666987631056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,5120,0.03635822070969476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,12288,0.10487555795245701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,16384,0.06565155585606892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,51200,0.06773600313398573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,16384,0.02299022177855174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,12288,0.08848000235027736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,3072,0.00814133303032981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,1024,0.020672889219390023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,32,0.003906666818592283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,12288,0.015581332974963717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,4096,0.028304888142479792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,3584,0.0772435532675849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,512,0.0034373334298531213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,6144,0.01999288962946998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,64,0.002575111058023241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,512,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,65536,0.3968666659461127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,6144,0.01924088928434584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,768,0.014752000570297241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,5120,0.010053333308961656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,2560,0.007274666594134436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,5120,0.044469333357281156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,32,0.015491555134455362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,7168,0.4770231246948242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,512,0.05537155601713392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,128,0.017252445220947266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,256,0.020243555307388306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,12288,0.026068444053332012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,51200,0.9513431125217013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,12288,0.022767111659049988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,6144,0.07749599880642362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,12288,1.7022337383694117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,64,0.006870222174459034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,1536,0.020920889245139226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,6144,0.008600888980759515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,768,0.09298221932517158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,65536,0.19910311698913574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,512,0.015486222174432544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,3072,0.014704000618722705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,128,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,16384,0.1084337764316135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,128,0.03294310967127482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,1024,0.009283555878533257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,10240,0.04533066683345371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,1536,0.010611555642551847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,4096,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,3584,0.010942222343550788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,3584,0.008216888540320927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,128,0.02789777848455641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,51200,0.5119048754374186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,512,0.00628266649113761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,3072,0.10164178080028957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,5120,0.012480888929631976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,2048,0.007318221860461765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,4096,0.009543110926946005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,3584,0.025352888637118872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,10240,0.3394160005781386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,51200,0.21657333109113905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,8192,0.02084000077512529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,256,0.013024888932704926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,3584,0.17046844959259033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,64,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,4096,1.4222506417168512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,6144,0.03830400109291077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,10240,0.49775112999810117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,8192,0.1597128841612074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,3072,0.09896088971032037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,6144,2.211141374376085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,8192,0.008788444101810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,512,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,3072,0.010120888551076254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,10240,0.01742488808102078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,65536,0.40410576926337344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,65536,0.15738577312893337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,2560,0.06974222262700398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,3072,0.006806222101052602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,6144,0.20240978399912515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,64,0.0053386667536364655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,512,0.004436444491147995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,5120,0.07720355855094062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,3584,0.2539190981123182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,128,0.0028026666906144884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,3072,0.006488889041874144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,3584,0.031139555904600356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,3072,0.03382044368320041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,12288,0.5528248680962456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,2560,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,51200,0.3298906750149197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,65536,0.37701956431070965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,7168,1.5735226737128363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,8192,0.025504888759719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,1536,0.0063635553750726916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,8192,0.013437333206335703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,7168,0.014828445182906257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,1024,0.023102222217453852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,8192,0.02146222194035848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,128,0.005145777844720417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,3584,0.06459911002053155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,6144,0.023538667294714186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,3584,0.014902222487661572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,2560,0.007126222054163615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,512,0.009703111317422655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,2560,0.08945155806011623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,3072,0.006264888991912206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,768,0.004242666479614046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,12288,0.019899555378490023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,32,0.009660444325870937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,32,0.03372711274358962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,16384,0.0507919987042745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,6144,0.017979555659823947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,4096,0.04807288779152764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,8192,0.022904000348514978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,2560,0.07957777712080213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,5120,0.1693706644905938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,2560,0.005690666536490123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,768,0.03387377659479777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,256,0.004955555415815777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,256,0.02751733362674713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,256,0.009175111022260454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,128,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,4096,0.014212444424629211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,2048,0.010118222071064843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,8192,0.04818400078349643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,3072,0.014546665880415173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,7168,0.04088622331619263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,2048,0.09169244766235352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,3072,0.008446222378147973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,512,0.004744889007674323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,10240,0.0069413334131240845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,4096,0.013800889253616333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,1024,0.27062400182088214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,1024,0.10813688569598728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,5120,0.009406222237481011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,8192,1.5642301771375868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,10240,0.3591199980841742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,8192,0.2620035542382134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,3584,0.010229333407349056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,5120,0.014344000154071383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,5120,0.006638222270541721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,6144,0.038423111041386925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,8192,0.0346995559003618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,2048,0.006720888945791457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,10240,0.02847111225128174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,51200,19.043482462565105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,10240,0.04507466819551256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,10240,0.085844443904029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,6144,0.32150043381585014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,768,0.03169155451986525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,51200,0.013616888887352414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,4096,0.02980177932315403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,256,0.03156355685657925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,1536,0.010647999743620554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,65536,0.07119644350475736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,128,0.0041511112617121804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,65536,0.564209779103597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,1536,0.006396444307433234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,32,0.0038328886859946777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,6144,0.0963475571738349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,5120,0.017346666918860543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,7168,0.14025421937306723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,512,0.009707555174827576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,3584,0.30439644389682347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,768,0.01404711107412974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,1536,0.006572444405820634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,3584,0.0989111132091946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,32,0.020095111595259774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,10240,0.02169422143035465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,128,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,3584,0.008838222258620793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,1024,0.006137777947717243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,65536,6.021823883056641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,12288,0.3214248816172282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,3072,0.030530667967266504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,7168,0.12177777290344238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,7168,0.07932177517149183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,3072,0.17793956067827013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,2048,0.009893333746327294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,2560,0.08729422092437744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,512,0.003294222263826264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,8192,0.01566844516330295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,8192,0.022650667362742957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,2048,0.0247999992635515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,2048,0.025399999486075506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,6144,0.024159111910396155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,51200,0.23484177059597441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,6144,0.018187556001875136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,65536,0.06846844487720065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,51200,1.3579902648925781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,2048,0.03814222084151374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,7168,0.013012444807423485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,32,0.007803555164072249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,2048,0.12583466370900473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,16384,0.22642932997809517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,16384,0.027246221899986267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,5120,0.03037422233157688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,2560,0.022201778160201177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,64,0.003042666655447748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,5120,0.022034666604465906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,12288,0.055135111014048256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,2560,0.013173333472675748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,1024,0.043149332205454506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,10240,0.02414488957987891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,10240,0.024992000725534227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,32,0.006425777657164468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,128,0.004294222427739037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,64,0.0024862223201327855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,3584,0.03397688931889004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,51200,0.4481786621941461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,512,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,256,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,32,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,10240,0.034083555142084755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,16384,0.0063946665161185795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,12288,0.03805422120624118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,128,0.071743110815684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,32,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,5120,0.0306275553173489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,51200,2.7971502939860025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,3584,0.010471999645233154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,2048,0.13984444406297472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,2560,0.006992888947327931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,10240,0.015328889091809591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,8192,0.021242666575643752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,51200,0.15160977840423584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,65536,0.18779467211829293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,5120,0.2495786613888211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,51200,0.05361510978804695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,65536,0.2714471022288005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,10240,0.14086666372087267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,12288,0.04246222310596042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,1536,0.08007555537753634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,2560,0.06797511047787137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,2560,0.042035556501812406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,12288,0.03117155697610643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,2048,0.1545999977323744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,2560,0.006794666664467917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,2048,0.12591289149390325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,3584,0.007311111523045435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,2560,0.009992889232105678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,5120,0.10949955383936565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,4096,0.024154666397306655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,32,0.002520888836847411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,1024,0.010098666780524785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,1024,0.014832889040311178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,512,0.23329244719611275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,65536,0.9083119498358833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,7168,0.006341333190600078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,10240,0.24708978335062662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,2048,0.02755911151568095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,1024,0.011979555918110741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,256,0.006477333191368315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,10240,1.5296701855129664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,64,0.00851288851764467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,5120,0.10077333450317383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,3072,0.06376977761586507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,256,0.0034382223255104492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,768,0.006467555546098285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,16384,0.00775911079512702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,5120,0.059634665648142494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,8192,0.0646373364660475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,65536,15.445793999565971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,3072,0.030455999904208716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,512,0.01684799955950843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,8192,0.17932444148593477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,8192,0.13111733065711126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,768,0.009454222189055549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,1024,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,2560,0.011893333660231696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,7168,0.024872000018755596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,16384,0.028100444210900202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,5120,0.061680886480543345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,10240,0.04039288891686334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,51200,0.06176355812284681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,65536,0.0870551135804918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,7168,0.21394133567810059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,3072,0.008088889221350351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,12288,0.01590399940808614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,8192,0.10335289107428657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,6144,0.021073778470357258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,6144,0.009343110852771336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,512,0.005324444423119227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,16384,0.7715999815199109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,4096,0.0491226679748959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,2560,0.016128889388508268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,2560,0.014122666584120857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,16384,0.16287822193569607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,8192,0.029448890023761328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,65536,2.197271982828776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,65536,1.3520818286471898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,512,0.005183111048407025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,2048,0.03684355484114753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,5120,0.010946667028797997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,2560,0.01921422282854716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,768,0.09269333548016018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,32,0.0025964443468385274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,2048,0.008815111385451423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,512,0.03026133444574144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,1536,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,768,0.00554577757914861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,3584,0.028239998552534316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,12288,0.056383113066355385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,768,0.05405422382884555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,512,0.008573333422342936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,32,0.006912888752089606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,6144,0.34412267473008895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,6144,0.5664941999647353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,768,0.004057777838574516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,2560,0.06808711422814263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,64,0.005176888985766305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,1024,0.2976062297821045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,8192,0.1252195570203993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,512,0.006839111033413146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,12288,0.02437688906987508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,4096,0.15159289042154947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,3584,0.03003022074699402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,65536,14.111404418945312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,6144,0.02256711158487532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,4096,0.10326488812764485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,256,0.19013955858018664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,4096,0.010963555839326648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,51200,0.10382311211691962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,8192,0.03952266772588094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,7168,0.05962044662899441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,3072,0.008340444829728868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,4096,2.9365047878689237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,16384,0.02973244587580363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,65536,0.11158488856421576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,16384,0.09102755784988403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,8192,0.020959999826219346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,4096,0.008626666333940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,7168,0.01959022217326694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,128,0.0028568889117903183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,3584,0.00655644428398874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,1024,0.009874666730562845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,16384,0.20101422733730742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,16384,0.021543999512990315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,12288,0.21866310967339408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,16384,0.044349332650502525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,32,0.012111999922328524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,6144,0.07910755607816908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,1024,0.010175111393133799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,3584,0.009457777771684859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,16384,0.3523671097225613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,512,0.005490666462315454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,8192,0.027207111318906147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,16384,0.02734755476315816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,3584,0.026350221700138513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,256,0.0057591112951437635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,768,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,51200,0.18411289321051705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,768,0.04898577928543091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,2048,0.02769688930776384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,32,0.0310515562693278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,7168,0.08283378018273248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,16384,0.03631822268168131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,3584,0.14384622044033474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,7168,0.04604977700445387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,128,0.01993777851263682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,8192,1.4285297393798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,5120,0.017424000634087455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,7168,0.06363466713163587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,1536,0.024559999505678814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,51200,0.42520978715684676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,768,0.023568888505299885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,6144,0.050026665131251015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,1024,0.019513777560657926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,64,0.0028435554769304064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,1536,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,51200,0.05801510810852051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,12288,0.06929600238800049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,6144,0.11542133490244548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,6144,0.013651555611027611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,512,0.02307555576165517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,5120,0.0124195557501581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,512,0.008328888979223039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,3072,0.08517422278722127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,65536,0.4434151119656033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,3072,0.012279999752839407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,51200,0.04907733201980591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,3584,0.03544800149069892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,3072,0.14602577686309814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,1024,0.005944889038801193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,5120,0.07328355312347412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,12288,0.038324445486068726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,32,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,512,0.008730666504965888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,64,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,4096,0.006160888820886612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,768,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,51200,1.7748266855875652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,6144,0.03325600094265408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,16384,0.022882666852739122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,65536,0.07866844203737047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,512,0.0052328887912962176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,16384,0.10558844274944729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,6144,0.01622488929165734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,51200,3.1842257181803384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,256,0.0036515556275844574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,64,0.013790222505728403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,2560,0.01981511049800449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,65536,0.12768533494737414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,51200,0.009760888914267222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,4096,0.008282666405042013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,64,0.005450666778617435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,16384,0.10273689031600952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,51200,0.37491289774576825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,256,0.013393777940008374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,768,0.005777777896987067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,64,0.003622222277853224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,768,0.021362667282422382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,3072,0.019837333096398246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,3584,0.08433689011467828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,12288,0.3712533315022786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,512,0.005068444543414646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,6144,0.0232631117105484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,65536,1.2266249126858182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,256,0.10234577788247003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,16384,0.03517511155870225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,6144,0.08101955387327406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,10240,0.0399893323580424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,128,0.009758222434255812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,3072,0.07772799995210436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,2048,0.006649777707126405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,3584,0.17227199342515734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,1024,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,1536,0.00848711116446389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,2560,0.05369955632421705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,7168,0.016008888681729633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,7168,0.016583111551072862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,2048,0.013256889250543384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,128,0.0028213332924577924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,256,0.00435644429590967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,12288,0.03356800145573086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,3584,0.02141688929663764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,8192,0.20311910576290557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,768,0.003604444364706675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,768,0.044963555203543774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,7168,0.01253600004646513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,256,0.01938311093383365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,3584,0.014295111099878946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,6144,0.025641777449183997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,3584,0.06507200002670288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,512,0.015493333339691162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,6144,0.0851751102341546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,1536,0.014984000060293408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,5120,0.014020444618331062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,5120,0.16230311658647326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,512,0.01129244433508979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,512,0.005948444621430502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,65536,0.5432266659206814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,4096,0.19736977418263754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,1024,0.003993777765168084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,5120,0.05373511049482557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,4096,0.033626665671666466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,32,0.010951110886202918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,1024,0.052489777406056724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,512,0.0053546664615472155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,2560,0.006093333164850871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,128,0.01221333361334271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,7168,0.043677333328458995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,10240,0.02095555596881443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,128,0.004655111167165968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,768,0.006146666490369373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,7168,0.017345777816242643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,1024,0.005864000154866113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,10240,0.013267555170589022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,512,0.040592888991038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,16384,1.1496995290120442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,4096,0.011241777903503843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,256,0.008655110994974772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,8192,0.062104887432522245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,16384,4.8671082390679254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,32,0.019762666689025033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,4096,0.014412444498803882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,128,1.5251732932196722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,256,0.01202666676706738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,16384,0.0698924462000529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,8192,0.5266213417053223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,16384,0.02898755669593811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,51200,0.8946915732489692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,512,0.008953777452309927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,256,0.004858666823969947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,2048,0.01788444485929277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,4096,0.03889244463708665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,256,0.005633777628342311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,2560,0.009538667069541084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,16384,0.13795110914442274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,1024,0.007860444486141205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,16384,0.14527289072672525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,512,0.005113777601056629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,64,0.007567110988828871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,1536,0.203256885210673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,65536,0.07518133189943102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,512,0.008615111311276754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,2048,0.08052000072267321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,2048,0.00776711106300354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,512,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,64,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,8192,0.7592515415615506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,4096,0.008288000192907121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,1536,0.005422222117582957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,1024,0.004008888784382078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,8192,0.01979466610484653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,3584,0.10136889086829291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,1536,0.03089955449104309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,51200,0.09196888738208348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,51200,0.6266658041212294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,256,0.007699555820888943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,64,0.014198222094111972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,12288,0.03680533170700073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,2560,0.008901333643330468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,16384,0.034671111239327326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,51200,0.044312887721591525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,128,0.0058853332367208265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,8192,0.030921777089436848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,128,0.29211555586920845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,10240,0.010372444159454769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,768,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,65536,1.056513786315918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,512,0.004638222356637319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,5120,0.061612447102864586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,5120,0.19448799557156035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,12288,0.2167688873079088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,8192,0.09666755464341904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,256,0.020410666863123577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,2048,0.00498488876554701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,7168,0.04883466826544868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,16384,0.0251893334918552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,16384,0.6427457597520616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,16384,1.7386435402764215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,6144,0.07409866650899251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,32,0.006676444576846228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,256,0.021503999829292297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,3584,0.023132445083724126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,5120,0.15298310915629068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,4096,0.049885332584381104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,8192,0.16308177842034235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,64,0.0027546667390399506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,7168,0.03697155581580268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,32,0.0026853332916895547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,12288,0.09439733293321396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,256,0.004581333448489507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,256,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,10240,0.011618666350841522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,768,0.006437333507670297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,2048,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,65536,0.1361315515306261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,10240,0.017290666699409485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,64,0.015241778559154935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,51200,0.11686400572458903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,128,0.01665866706106398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,768,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,512,0.10408711433410645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,768,0.05693155527114868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,128,0.008190222084522247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,1536,0.004724444614516364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,2048,0.006978666616810693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,64,0.009514666265911525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,10240,0.006521777974234686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,256,0.16170488463507757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,16384,0.1981724500656128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,512,0.015277332729763456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,51200,1.0268932978312175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,65536,1.928696102566189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,64,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,64,0.027629334065649245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,512,0.005505777895450592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,12288,0.10612533489863078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,6144,0.14390133486853704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,8192,0.03179288903872172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,51200,0.08185155524147882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,65536,0.2461200025346544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,1024,0.017048888736300997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,4096,0.022387555903858606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,768,0.09086133374108209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,65536,0.07634577486250135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,12288,3.53351190355089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,4096,0.12308443917168511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,65536,1.7386070887247722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,3584,0.11295999421013726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,1024,0.006681777950790193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,512,0.02251288957066006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,5120,0.01868799991077847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,65536,1.788122600979275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,1024,0.00775999989774492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,16384,0.12967822286817762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,768,0.017993777990341187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,4096,0.057242665025922984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,1024,0.0529075562953949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,10240,0.2660551071166992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,16384,0.5178204642401801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,10240,0.717460420396593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,12288,0.048166222042507596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,768,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,128,0.012243555651770698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,32,0.017699556218253244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,2048,0.0974044468667772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,64,0.006654222392373615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,1024,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,4096,0.11065422164069282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,51200,0.05418311225043403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,5120,0.013444444371594323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,32,0.002963555562827322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,2560,0.024879111184014216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,5120,0.028412444723976985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,10240,0.0589582191573249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,3584,0.0068408888247278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,1024,0.009202666580677032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,8192,0.015788444214397006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,3584,3.089610629611545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,256,0.007697777615653143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,16384,0.0718924469417996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,5120,0.09163466427061294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,128,0.0069822221994400024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,512,0.015141333142916361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,51200,2.9037289089626737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,512,0.010446222292052375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,256,0.017478222648302715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,4096,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,16384,0.07129777802361383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,4096,0.0657128890355428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,5120,0.01072533345884747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,512,0.005782222168313132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,12288,0.4339919884999593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,6144,0.0860782199435764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,128,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,16384,0.07649066713121203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,512,0.0053191110491752625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,2560,0.04007110993067423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,65536,0.09822577900356716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,2048,0.006613333192136552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,2048,0.08773955371644761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,7168,0.0119982221060329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,8192,0.014299554957283868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,2048,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,4096,0.01273333364062839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,6144,1.2834355036417644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,128,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,5120,0.02819377846188015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,768,0.004488888714048597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,10240,0.266503996319241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,5120,0.011074666347768573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,10240,0.0473235547542572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,32,0.043807999955283276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,7168,0.01535555554760827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,1536,0.011328000161382886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,8192,0.028138667345046997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,3072,0.012383999923865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,768,0.06731200218200684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,768,0.005872888697518243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,12288,0.7102470927768283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,512,0.029236445824305218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,1024,0.010211555494202508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,4096,0.011189333266682096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,16384,0.03994933433002896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,65536,0.48708179261949325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,1024,0.005399111244413588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,10240,0.5855937533908421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,5120,0.015303111738628812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,16384,0.6726151042514377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,3072,0.010033778018421596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,3584,0.036372443040212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,256,0.004090666770935059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,3072,0.10197955369949341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,1024,0.030982222821977403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,6144,0.056424889299604625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,128,0.004833777745564778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,256,0.006214222146405115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,10240,5.17994647555881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,6144,0.0912968913714091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,64,0.00361333332127995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,3072,0.00814844419558843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,128,0.017860444055663217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,768,0.5626977814568414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,7168,0.02948088778389825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,2560,0.1065937810473972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,32,0.004100444416205089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,1024,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,2560,0.008180444439252218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,5120,0.4311973253885905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,5120,0.014378666877746582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,4096,0.0563288860850864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,3584,0.07135022348827787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,7168,0.02460533380508423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,1536,0.009816889133718278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,512,0.00535999983549118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,2048,0.005584888988071018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,512,0.03923022084765964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,32,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,51200,0.2842160065968831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,1536,0.006877333339717653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,3584,0.015800888339678448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,1536,0.026504000027974445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,65536,0.06794044706556532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,3584,0.31144266658359104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,10240,0.04336622357368469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,512,0.04971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,5120,0.013799111048380533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,16384,1.0271049075656467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,65536,0.01770400007565816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,6144,0.02574488851759169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,5120,0.03399822115898132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,1024,0.027334221535258826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,8192,0.015463999576038785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,65536,0.27731111314561635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,1024,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,768,0.05752978059980604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,65536,0.11398933993445502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,3584,0.010255999863147736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,6144,0.04904355605443319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,64,0.004848000076082018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,8192,0.15630666414896646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,3584,0.05857866340213352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,128,0.0031031111462248694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,1536,0.6677350997924805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,7168,0.015415999624464246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,65536,0.5410444471571181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,3584,0.0094035557574696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,8192,0.23783198992411295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,16384,0.015967110792795818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,2048,0.005099555684460534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,512,0.022505778405401442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,3584,0.011928888658682505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,512,0.0034826666944556763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,3584,0.032039112514919706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,10240,0.10706933339436848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,256,0.005341333233647876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,10240,0.03752800159984165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,8192,0.024329778220918443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,1024,0.02610133257177141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,64,0.0029644444584846497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,65536,0.18774133258395723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,10240,2.474494298299154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,32,0.010592888626787396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,51200,0.18452444341447619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,32,0.03706311186154684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,7168,4.045035468207465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,8192,0.03792622354295518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,512,0.2536897659301758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,768,0.014134221606784396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,6144,0.8142124281989204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,32,0.008172444171375698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,768,0.10229066345426772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,3072,0.00812977800766627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,256,0.30613867441813153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,6144,0.016873778568373788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,16384,0.043883555465274386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,3072,0.14714488718244764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,4096,0.022487110561794702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,1536,0.034155554241604276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,768,0.016005333926942613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,256,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,3072,0.0266275554895401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,64,0.00433688900536961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,5120,0.018817777434984844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,64,0.005928888916969299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,8192,0.1216391060087416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,51200,0.12792266739739314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,6144,0.15843821896447077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,6144,0.4292906655205621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,1024,0.0068853331936730286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,2560,0.031713777118259005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,51200,1.7828559875488281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,8192,0.14805511633555093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,256,0.0035786665976047516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,10240,0.03695822093221877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,2048,0.041533334387673274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,10240,0.02795733345879449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,1024,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,3072,0.037652442852656044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,12288,0.38039999537997776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,65536,0.13100355201297337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,8192,0.03811555438571506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,10240,0.16421243879530165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,512,0.030995554394192163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,5120,2.198663075764974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,10240,0.03184444374508328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,2048,0.010781333678298526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,128,0.14788799815707737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,16384,0.006805333412355847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,32,0.024700444605615404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,3072,0.013639111485746173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,4096,0.09493777487013076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,2048,0.008806222014957005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,65536,0.2705119980706109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,3584,0.010584000084135266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,10240,0.01741333305835724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,65536,0.22957865397135416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,16384,0.039432889885372586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,8192,0.06977599859237671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,16384,0.3753884368472629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,51200,0.12501600053575304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,4096,0.01368799971209632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,32,0.0028515555378463534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,2048,0.01293777757220798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,5120,0.010465777582592435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,128,0.004530666602982415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,16384,0.11927378177642822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,51200,0.9573716057671441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,1024,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,16384,0.5982266532050239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,2048,0.0068400001360310455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,64,0.010950222611427307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,65536,0.08366666899787055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,768,0.019609777463806998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,512,0.007189333438873291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,16384,0.05150755577617221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,2048,0.9709217283460828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,1024,0.6261546876695421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,65536,0.07861510912577312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,6144,0.014634667171372307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,768,0.029733333322736952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,2560,0.009566222627957663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,1024,0.11908533838060166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,8192,0.0306915541489919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,12288,0.020043555233213637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,16384,0.9315964380900065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,768,0.017588444881969027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,6144,0.017510222064124215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,10240,0.025983999172846477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,5120,0.01626311077011956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,1024,0.007311111523045435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,7168,0.15992355346679688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,1536,0.05752444267272949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,8192,0.006824888702895906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,65536,0.39324355125427246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,65536,0.1454577843348185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,3584,0.01036266651418474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,5120,0.02342044479317135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,512,0.004726222405831019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,51200,0.04750400119357639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,6144,0.06690044535530938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,6144,0.007195555501514011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,768,0.024148444334665935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,16384,0.07575022512012057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,3584,0.153949326939053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,7168,0.011347555451922946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,51200,0.11175999376508926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,3584,0.1792293389638265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,7168,0.15948000219133165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,4096,0.006022222340106964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,7168,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,512,0.00721599989467197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,51200,0.051311999559402466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,5120,0.009287111461162567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,5120,0.37920088238186306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,5120,0.007353777686754863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,2560,0.4040035671657986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,12288,0.10146844387054443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,10240,0.3937413427564833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,256,0.047480000389946826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,3584,0.010203555226325989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,65536,0.8161919911702474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,4096,0.08867288960350884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,128,0.015444444285498725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,128,0.0026453334010309647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,1536,0.014304889572991265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,1536,0.06848888927035861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,65536,0.1704826619890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,2560,0.16929954952663848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,8192,0.05962933434380425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,512,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,51200,0.9823652903238932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,51200,0.3744355572594537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,51200,2.3231707678900824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,7168,0.10770844750934178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,64,0.00810222244924969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,32,0.007528000407748752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,8192,0.015021334091822306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,12288,0.01109955542617374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,128,0.003237333355678452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,65536,2.2702852884928384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,1536,0.018248889181349013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,7168,0.26066221131218803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,51200,0.010792888700962067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,2560,0.008457777400811514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,51200,1.8504852718777125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,1536,0.008057778080304464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,7168,0.22333865695529512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,3584,0.05452444487147861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,128,0.006179555422729916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,2560,0.016536888149049547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,8192,0.10224711232715183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,65536,0.2394204404619005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,8192,0.052856001589033336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,10240,0.0355022218492296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,2048,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,32,0.025059555967648823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,3584,0.12008799446953668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,1536,0.027777777777777776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,51200,0.13222400347391763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,64,0.013839110732078552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,2560,0.007134222322040134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,65536,1.0019493103027344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,2560,0.0075839997993575195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,5120,0.06854577859242757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,8192,0.031585776143603854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,2048,0.13340977827707926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,3584,0.057465778456793896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,1536,0.006698666761318843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,128,0.008831111093362173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,4096,0.005980444451173146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,7168,0.049954666031731494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,1024,0.0037271111375755737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,64,0.0034053333931499054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,5120,0.021028444170951843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,768,0.007861333588759104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,32,0.0030177777840031516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,512,0.014896889527638754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,5120,0.00776977754301495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,64,0.03570577833387587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,8192,0.016116445263226826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,6144,0.01149777736928728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,2048,0.01033866653839747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,16384,0.07688799831602308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,16384,0.02092444399992625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,6144,0.2540515528784858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,64,0.00546488869521353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,1536,0.010453333457310995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,7168,0.38588889439900714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,2560,0.011580444044537015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,3584,0.05064444409476387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,1024,0.020417778028382193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,8192,0.029412445094850328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,65536,1.25872802734375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,2048,0.03383200036154853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,5120,0.23601865768432617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,128,0.023806222611003455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,32,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,2048,0.007027555671003129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,6144,0.055530667304992676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,128,0.00442844463719262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,10240,0.29474221335517037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,7168,0.008256889051861232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,128,0.0025600000388092464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,10240,1.9489555358886719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,8192,0.5746222072177464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,768,0.2698231008317736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,2560,0.8244826528761121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,4096,0.015991111596425373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,12288,0.03298755486806234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,4096,0.011580444044537015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,12288,0.038112001286612615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,2560,0.00870222184393141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,2048,0.010989333192507425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,5120,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,768,0.016138666205936007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,65536,0.06903200017081366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,10240,0.013717333475748697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,2560,0.008461332983440822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,3584,0.06393777661853366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,128,0.20671466986338297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,256,0.005333333379692501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,2048,0.012079111403889127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,2560,0.007916444705592262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,2560,0.012088889049159156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,3584,0.012421333127551608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,6144,0.006214222146405115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,10240,0.4159671200646295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,10240,0.01633777717749278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,2560,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,3072,0.06432977649900648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,5120,0.01015911085738076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,4096,0.05297866794798109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,64,0.0028453332682450614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,8192,0.024150222539901733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,7168,0.116138670179579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,32,0.0028497777465316984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,16384,0.03507288959291246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,10240,0.020579554968410067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,64,0.002855111120475663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,65536,6.052726321750217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,64,0.022111111217074927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,256,0.003370666669474708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,32,0.010631110933091907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,65536,0.4365075429280599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,64,0.003519999898142285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,6144,0.658390204111735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,2048,0.03792533278465271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,128,0.005063111169470681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,1024,0.016897777716318767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,10240,0.018405333161354065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,16384,0.03381333417362637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,5120,0.19802755779690215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,2048,0.12163644366794163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,5120,3.039744906955295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,512,0.00869688888390859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,1536,0.014688889185587565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,65536,1.8718107011583116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,2048,0.03465777635574341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,12288,0.12015377150641547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,8192,0.04836000005404154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,128,0.010693333215183683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,4096,0.009150221943855286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,16384,0.07486044698291354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,51200,0.8030257754855685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,4096,0.011911110745535957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,32,0.014754666222466363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,3072,0.0709386666615804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,1536,0.006849777781301075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,3584,0.028831998507181805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,65536,15.589502970377604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,51200,0.07360711362626818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,2048,0.007373332977294922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,1024,0.007648000286685095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,4096,0.01633777717749278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,3584,0.07058488660388522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,16384,0.04153866569201151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,768,0.005977777971161737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,32,0.0053546664615472155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,3072,0.00943911075592041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,2048,0.009351111120647853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,6144,0.013264888690577613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,2048,0.06577333476808336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,1024,0.004038222134113312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,2048,0.0058355554938316345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,32,0.01204355557759603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,768,0.008237332933478886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,3072,0.013286222186353473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,3584,0.22698044776916504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,65536,0.19967199696434867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,16384,0.03905333413018121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,16384,0.05937777625189888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,768,0.004578666554556953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,3584,1.3376204172770183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,256,0.0041404445138242514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,3584,1.0374310811360676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,1024,0.29134755664401585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,2048,0.181532449192471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,128,0.008764444126023186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,128,0.004963555683692296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,3072,0.025752888785468206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,512,0.008006222546100616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,5120,0.057929780748155385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,512,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,64,0.0026391111314296722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,16384,0.024530667397711012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,1024,0.007373332977294922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,16384,0.04262400004598829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,3584,0.015224888920783997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,1536,0.01256711118751102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,51200,0.050937778419918485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,51200,0.09958044687906902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,5120,0.33594756656222874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,3072,0.11770578225453694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,12288,0.012072000238630505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,3584,0.03774844606717428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,7168,0.3887493345472548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,8192,0.021511110994550917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,256,0.04862755537033081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,512,0.005613333235184352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,16384,0.21410489082336426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,7168,0.010509333676762052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,128,0.0053075556125905775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,1024,0.02791822287771437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,3584,0.16054577297634548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,32,0.0041982221106688184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,12288,0.05603555838267008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,5120,0.00981777740849389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,32,0.005917333480384614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,256,0.003055111194650332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,3584,0.05071022113164266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,2048,0.011100444528791638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,512,0.011583111352390714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,16384,0.12681867016686335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,65536,0.06549600097868177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,7168,0.01814933286772834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,512,0.003560888684458203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,2560,0.14508888456556532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,16384,0.03626577721701728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,10240,0.059359113375345864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,3584,0.01519733336236742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,1024,0.011672000090281168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,7168,0.34683820936414933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,512,0.06897333595487806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,2560,0.008522666990756989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,64,0.006881777611043718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,6144,0.019253333409627277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,51200,0.12936977545420328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,3072,0.028838220569822524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,768,0.02089244458410475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,8192,0.043974220752716064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,512,0.005611555443869696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,8192,0.025497777594460383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,3584,0.06766755713356866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,32,0.003375111147761345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,4096,0.2726648913489448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,65536,1.2066222296820748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,4096,0.08199822240405612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,4096,0.0151555554734336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,65536,0.07158844338523017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,8192,0.03527199890878465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,16384,0.049492445256974965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,12288,0.3262337843577067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,3072,0.009030222064918941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,6144,0.137807117568122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,5120,0.044481777482562594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,12288,0.02681244413057963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,51200,0.06203555398517185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,16384,0.022474666436513264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,1024,0.006562666760550604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,3584,0.02422844370206197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,3584,0.01349688900841607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,1536,0.027441778116756018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,64,0.10138577885097927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,2048,0.013929777675204806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,10240,0.06891733407974243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,2560,0.007182222273614671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,3072,0.05840888950559828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,10240,0.028424888849258423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,16384,0.540795538160536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,768,0.005030222237110138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,768,0.008482666479216682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,768,0.011237333218256632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,7168,0.5290364689297146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,16384,0.12861955165863037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,7168,0.04090399874581231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,3072,0.019848888119061787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,12288,0.07593599955240886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,3072,0.026173333326975506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,4096,0.07764977878994413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,3072,0.061866667535569936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,64,0.003439999909864532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,16384,0.15672443972693548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,2560,0.13056000073750815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,128,0.0026986667265494666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,5120,0.3213591045803494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,10240,7.273164537217881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,1024,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,2048,0.12326399485270183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,8192,0.01657511128319634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,16384,0.049638221661249794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,1024,0.4303920004102919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,2560,0.03735288977622986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,128,0.005523555394675996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,128,0.00480266660451889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,10240,0.26204000578986275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,512,0.03642844491534763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,256,0.01702311138312022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,6144,0.026779555612140234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,16384,0.021360889077186584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,128,0.004931555440028508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,32,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,65536,0.2783466709984673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,512,0.020477333002620272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,10240,0.1709057754940457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,8192,0.03155022197299533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,1536,0.013839110732078552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,8192,0.05349511239263746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,256,0.02350666622320811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,768,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,3072,0.07596088780297174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,10240,3.892918904622396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,3072,0.058283554183112256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,16384,0.0859066645304362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,4096,0.0888035562303331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,2048,0.10683644480175442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,4096,0.03862399856249491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,6144,0.01146577795346578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,768,0.28286221292283803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,64,0.0027395555128653846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,128,0.004871110949251386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,10240,0.33039644029405385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,768,0.01369955556260215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,6144,0.015262222952312894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,512,0.015607999430762397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,2048,0.6080506642659506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,64,0.013559111290507846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,3072,0.20536533991495767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,3072,0.05363466673427158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,6144,1.0440222422281902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,16384,0.4553866916232639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,256,0.0032951111594835916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,7168,0.023521777656343248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,16384,0.504250685373942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,7168,0.07705599731869169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,1024,0.006299555715587404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,2560,0.035206221871905856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,16384,0.24030399322509766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,32,0.019674667053752475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,768,0.019455111689037748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,7168,0.01633066601223416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,8192,0.026407111022207472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,768,0.07811199956470065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,65536,1.5763875113593206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,16384,0.047156443198521934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,5120,0.012535110943847232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,51200,0.09596533245510525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,768,0.018028444714016385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,16384,0.2512693405151367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,768,0.029861334297392104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,1024,0.034965333011415266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,65536,1.8402790493435328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,12288,0.0563031103875902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,6144,0.017222222354676988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,256,0.004859555512666702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,1024,0.1275120046403673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,6144,0.17762666278415254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,1024,0.3446648915608724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,256,0.009026666482289633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,3072,0.009382222261693742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,65536,13.165952894422743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,3072,0.026682666606373254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,51200,0.19146222538418242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,12288,0.02429955535464817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,10240,0.032207111517588295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,64,0.005295111073387994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,12288,0.01590133375591702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,5120,0.04555199874771965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,3584,0.012054222325483957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,5120,0.021241777473025854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,10240,0.037392887804243304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,3072,0.006767999794748094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,6144,0.01573866605758667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,5120,0.013397333522637686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,4096,0.12023022439744736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,1024,0.018519111805491976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,768,0.01849422189924452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,65536,0.34598933325873477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,1536,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,32,0.014723555909262763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,4096,0.030410667260487873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,12288,0.033258666594823204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,1024,0.014616888430383472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,256,0.01350044459104538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,16384,0.6472915543450249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,3584,0.09654577573140462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,16384,0.039142221212387085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,51200,4.744391123453776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,10240,0.01644177734851837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,3584,0.01165688865714603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,32,0.03996444410747952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,512,0.006565333240562015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,2048,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,6144,0.016131555040677387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,8192,0.013487111363146039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,8192,0.045066667927636035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,1536,0.004709333181381226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,2048,0.007442666424645319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,8192,0.1043342219458686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,4096,0.008345777789751688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,32,0.011253332926167382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,12288,0.09984978040059407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,512,0.008999110923873054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,65536,1.8843421936035156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,16384,0.030293332205878362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,51200,0.5410293473137749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,32,0.00370755543311437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,768,0.0203484445810318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,128,0.0033191111352708605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,128,0.003072000005178981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,10240,0.016358221570650738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,1024,0.008815999660227034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,8192,0.4501795768737793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,7168,0.09799555275175308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,1024,0.030446221431096394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,32,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,64,0.0075493330756823225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,1536,0.010789333118332757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,3072,0.07558221949471368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,8192,0.012210666305489011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,256,0.004600000050332812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,1536,0.01774311065673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,65536,2.0548345777723527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,128,0.005777777896987067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,4096,0.012700444294346703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,10240,1.8944435119628906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,65536,17.501230875651043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,512,0.05822044610977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,2048,0.10407911406622992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,4096,0.08623733123143514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,12288,0.018147556318177115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,2048,0.06592177682452731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,3584,0.0498560004764133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,256,0.011378666592968835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,8192,0.02049511174360911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,2048,0.05551200111707052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,51200,98.39716932508681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,51200,0.13694666491614446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,64,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,128,0.005065777649482091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,256,0.023984889189402264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,16384,0.024812445044517517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,1536,0.06061244673199124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,256,0.0035822221802340615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,2560,0.12117689185672337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,51200,0.19673066669040254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,12288,0.3406284385257297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,768,0.01923288901646932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,12288,0.006395555618736479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,512,0.31393954488966197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,10240,0.21855378150939941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,512,0.54676267835829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,32,0.014905777242448596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,7168,0.07789155509736803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,768,0.02520622147454156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,65536,0.7816061973571777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,4096,0.014409777190950183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,128,0.0053084443012873335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,768,0.007851555943489075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,6144,0.008288000192907121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,1024,0.06541155444251166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,1536,0.02561155623859829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,512,0.020006222857369315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,4096,0.9945715798272027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,64,0.0033359999457995095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,12288,0.09193066755930583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,51200,0.7950968742370605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,512,0.005904888941182031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,512,0.004424889054563311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,65536,0.1704559988445706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,2560,0.010263111028406356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,8192,0.4105457729763455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,1536,0.012838222086429596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,12288,0.06760266754362318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,1024,0.03953333364592658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,16384,0.011015999648306104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,128,0.035051554441452026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,3584,0.007259555160999298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,6144,0.11412889427608913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,2560,0.024375999967257183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,51200,1.1559333801269531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,8192,0.32196177376641166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,32,0.018948444061809115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,4096,0.011324444578753578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,1536,0.8121093114217123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,2560,0.8382195366753472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,2048,0.007376888559924231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,7168,0.02096533278624217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,7168,0.03157688842879401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,10240,0.023383110761642456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,768,0.004211555752489301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,7168,0.01202311118443807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,768,0.052799999713897705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,2048,0.0058906666106647914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,2560,0.09804355435901219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,128,0.002817777709828483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,1536,0.04615199897024366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,128,0.04655111167165968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,16384,1.0770177841186523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,16384,0.2445333268907335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,3072,0.010437332921557956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,16384,0.3956124517652724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,10240,0.020637333393096924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,16384,0.17604622575971815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,65536,0.12341954973008896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,2048,0.10931199789047241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,7168,0.18005243937174478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,4096,0.10268533229827881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,768,0.014448000325096978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,65536,0.0178195552693473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,5120,0.014780445231331719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,4096,0.054583999845716685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,32,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,1024,0.009203555683294931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,512,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,5120,0.022358222140206233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,3072,0.21869333585103354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,2048,0.029952890343136255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,1536,0.009659555223253038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,5120,0.07080888748168945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,12288,1.0740248362223308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,5120,0.007479999628331926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,4096,1.4999003940158422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,3584,0.01092266705301073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,3072,1.5280266867743597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,6144,0.013329777452680798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,512,0.0041857775714662345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,2048,0.029002666473388672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,5120,0.024266666836208765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,2560,0.373074664009942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,65536,0.24180889129638672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,512,0.04827288786570231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,8192,0.8778479894002279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,5120,0.009307555854320526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,3072,0.1808177762561374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,1024,0.007962666451931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,51200,1.3226808971828885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,10240,0.14715111255645752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,7168,0.013991110854678683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,8192,0.3371688789791531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,6144,0.024321777953041926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,64,0.0029324444217814338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,7168,3.252748489379883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,16384,0.8276764551798502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,2560,0.01531733406914605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,8192,0.2697849008772108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,7168,0.012986666626400419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,1024,0.01644177734851837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,10240,0.039643555879592896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,65536,0.060786664485931396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,6144,0.04784977767202589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,2048,0.010427555276287926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,3584,0.017163554827372234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,10240,0.5728569030761719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,512,0.009036444127559662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,12288,0.040363556808895536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,2560,0.008636444807052612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,1024,0.03471466567781236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,3584,0.009952888720565373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,512,0.008624888956546783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,1024,0.038507554266187884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,51200,2.6769510904947915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,16384,0.03993599944644504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,51200,0.585610654619005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,768,0.013795555465751223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,256,0.0038035557501845886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,256,0.005163555757866966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,7168,0.812568876478407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,1024,0.024274667104085285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,768,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,10240,0.006455111006895701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,2560,0.010017777482668558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,8192,0.011600000162919363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,768,0.10856978098551433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,4096,0.013013333082199097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,5120,0.03467466764979892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,256,0.010800888968838586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,4096,0.2597208817799886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,128,0.007080888582600488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,2048,0.8755751185946994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,1024,0.05716266896989611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,4096,0.026480888326962788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,65536,3.1179724799262156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,65536,0.06534577740563287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,768,0.0055440002017551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,7168,0.16415022479163274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,32,0.0027395555128653846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,16384,0.06231821907891167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,4096,0.11801777945624457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,6144,0.011852444046073489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,3584,0.006335111127959357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,1536,0.4352586534288194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,768,0.016787555482652452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,16384,0.023805333508385554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,32,0.0026702222724755607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,32,0.0027520000520679685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,128,0.005158222383923001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,64,0.018049778209792245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,1536,0.03194844391610887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,5120,0.014270222849316068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,10240,0.26928355958726674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,64,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,10240,0.18049155341254342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,5120,0.3874062167273627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,128,0.0037288889288902283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,7168,0.032237334383858576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,10240,0.040488001373079084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,512,0.009443555441167619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,65536,0.04604444570011563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,1024,0.11277243826124404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,12288,0.028051555156707764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,256,0.06541688574684991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,64,0.030088888274298772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,768,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,2560,0.015722667177518208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,2048,0.3990613354576959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,128,0.0068915556702348925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,4096,0.03265955713060167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,256,0.011586666935020022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,10240,0.052024887667761914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,65536,1.3625031577216253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,1024,0.007377777662542131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,3072,0.0602782236205207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,128,0.004122666600677702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,8192,0.01856355534659492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,6144,0.01240355521440506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,3584,0.05624000231424967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,512,0.007522666619883643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,1024,0.005481777919663324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,7168,0.039990223116344877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,256,0.15958933035532633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,256,0.04789955417315165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,65536,2.890138626098633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,128,0.0053297777970631914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,6144,0.011391110718250275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,3072,0.011123555401961008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,51200,0.06616089079115126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,512,0.01812088820669386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,1024,0.009286222358544668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,32,0.005151111218664381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,3584,0.030388444662094116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,2560,0.14757243792215982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,4096,0.03266577919324239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,64,0.002597333242495855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,51200,0.08665511343214248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,51200,0.16450933615366617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,1024,0.012355555262830524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,3072,0.013898667362001208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,1024,0.004053333153327306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,8192,0.018048889107174344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,7168,0.05715733104281955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,51200,1.665053367614746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,7168,0.20229155487484404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,4096,0.01651822196112739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,5120,0.03150044547186957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,3584,0.011013333168294696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,5120,0.023896000451511804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,3584,0.07795466979344685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,65536,0.2510044574737549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,7168,0.015881778465376962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,256,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,65536,8.467258877224392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,512,0.025195555554495916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,51200,0.18810310628679064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,10240,4.114305708143446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,2560,0.01551999979548984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,768,0.0070293330483966405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,768,0.027694221999910142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,7168,0.9309760199652778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,768,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,5120,0.20008711020151773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,51200,0.045610666275024414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,768,0.006509333435032103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,6144,0.03224444389343262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,12288,0.11781866020626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,12288,0.9277546140882703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,51200,0.15356711546579996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,12288,2.475757386949327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,65536,0.06955111026763916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,32,0.032039112514919706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,16384,0.041998220814598926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,2560,0.029986666308508977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,5120,0.1664186716079712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,12288,0.06208178069856432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,256,0.005341333233647876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,12288,0.028241776757770117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,1024,0.01756622228357527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,12288,0.10082933637830947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,16384,0.10966310898462932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,64,0.002586666701568498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,7168,0.014427555931939019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,10240,0.048933333820766874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,512,0.003777777983082665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,7168,0.3254693349202474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,512,0.011694221860832639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,2048,0.07999822166230944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,4096,0.014900444282425774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,10240,0.06547022528118558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,16384,0.04544622368282742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,6144,0.01071999967098236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,2560,0.010039110978444418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,1024,0.05204266640875074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,2560,0.06013155645794339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,6144,0.04307022359636095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,12288,0.02509511179394192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,512,0.02349777850839827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,1024,0.02126755482620663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,1536,0.03516711129082574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,256,0.0075795559419526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,64,0.0026844443960322272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,1024,0.5621173116895888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,7168,0.19689777162339953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,10240,0.31296179029676646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,32,0.005493333356248007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,256,0.18108710977766249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,3584,0.011201777391963534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,1536,0.03689422210057577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,2560,0.029672887590196397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,16384,0.4685217539469401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,6144,0.15062577194637722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,32,0.0026373333401150177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,2560,0.007237333390447829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,512,0.00683111117945777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,2560,0.04294844468434652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,256,0.0033573332346147964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,6144,0.029905776182810467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,3072,0.20119467046525744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,7168,0.014365333649847241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,64,0.00350222239891688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,128,0.0028515555378463534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,51200,0.7562284469604492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,256,0.013273778061072031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,8192,0.283881770239936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,2048,0.05248444279034933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,4096,0.023870221442646448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,51200,0.058372444576687284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,4096,0.05657244390911526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,6144,0.21789333555433485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,8192,0.038487109873029925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,16384,0.07010577784644233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,2048,0.013198222551080914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,128,0.008558221989207799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,768,0.005609777652555042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,10240,0.028599109914567735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,7168,0.13324177265167236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,16384,0.05980000231001112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,64,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,2560,0.09101066986719768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,6144,0.20449688699510363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,12288,0.04085777865515815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,2560,0.010812444819344414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,256,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,12288,0.19437066713968912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,4096,0.05436177717314827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,2560,0.032771554258134626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,128,0.004295111116435793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,512,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,2560,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,768,0.018373333745532565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,768,0.00832177781396442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,128,0.004015111260943943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,4096,0.1223742167154948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,256,0.008419555922349295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,6144,0.04276177618238661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,3584,0.05537688732147217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,16384,0.17379821671379936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,12288,0.012496888637542725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,51200,0.9832969241672092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,1024,0.04232622186342875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,6144,0.5517706871032715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,512,0.004293333325121137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,1536,0.007700444095664554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,3072,0.008546666966544257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,3584,0.03457866774664985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,7168,0.07607822285758124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,3584,0.01443911095460256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,2048,0.008782222039169734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,6144,0.016415999995337594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,4096,0.1825502183702257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,1024,0.017283555534150865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,128,0.021236444513003033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,2560,0.0632622241973877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,128,0.002959111084540685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,4096,0.01236088905069563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,3584,0.010013333625263637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,12288,0.538008901807997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,51200,1.0887271033393013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,5120,0.016170667277442086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,256,0.009358222285906473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,768,0.004763555609517627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,5120,0.1726435555352105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,3584,0.11733067035675049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,256,0.005879999862776862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,4096,0.12596089310116237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,65536,11.690582275390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,1536,0.007605333295133378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,3072,0.01706133286158244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,2048,0.005274666680230035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,256,0.007542221910423703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,4096,0.020658666888872784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,4096,0.1078195571899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,7168,0.2401075628068712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,8192,0.44429956542121035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,4096,0.01744622157679664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,6144,2.590837266710069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,1536,0.02055377761522929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,3584,0.01833066675398085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,768,0.004421333471934001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,128,0.003436444534195794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,512,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,16384,0.1281013356314765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,2048,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,6144,0.028088887532552082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,65536,0.13594310813479954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,6144,0.007792000141408708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,7168,0.09903733597861396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,768,0.006779555645253923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,4096,0.019517333971129525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,512,0.02327377763059404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,12288,0.04036888811323378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,32,0.018930666976504855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,256,0.004938666605287128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,4096,0.034407999780442976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,12288,0.061089780595567494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,2048,0.013825777504179211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,51200,0.03807288739416335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,1024,0.004314666820896996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,4096,0.02608888844648997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,65536,1.0166959762573242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,512,0.01982933282852173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,65536,11.502117580837675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,32,0.024622221787770588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,32,0.07779288954204984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,6144,0.03769244419203864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,128,0.21907911035749647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,5120,0.013340444200568728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,16384,5.305255889892578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,1536,0.03109244505564372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,8192,0.021288888321982488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,65536,0.07206577724880643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,51200,0.0843964417775472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,51200,3.9449128044976125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,2048,0.07515288723839654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,3072,0.25416623221503365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,16384,0.01594311164485084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,128,0.006599111275540457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,128,0.577730655670166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,2560,0.005602666487296422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,12288,0.0162808895111084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,51200,0.10693155394660102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,1536,0.012187555432319641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,768,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,128,0.004505777938498391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,2048,0.02520977788501316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,7168,0.019889778561062284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,8192,0.08418666654162937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,256,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,1024,0.03029777937465244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,1536,0.01240533341964086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,128,0.014464888307783337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,7168,0.02184088859293196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,4096,0.011960888902346293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,1536,0.010293333066834344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,8192,0.03412266572316488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,51200,0.22834844059414336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,128,0.0163093325164583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,3584,0.009855111440022787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,6144,0.01647911138004727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,64,0.029858665333853826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,256,0.018457776970333524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,1024,0.4383644527859158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,512,0.006057777752478917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,12288,0.021003555920388963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,256,0.006487111250559489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,3584,0.06366755565007527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,7168,0.04203288753827413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,51200,2.5061617957221136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,8192,0.13633510801527235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,2048,0.016168889072206285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,4096,0.028063999281989202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,512,0.005646222167544895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,5120,0.4374577734205458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,12288,0.04859288864665561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,64,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,12288,0.45042843288845485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,3072,0.012317332956526013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,3072,0.022484444909625586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,6144,0.041626665327284075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,51200,0.1326506667666965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,768,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,65536,0.14093511634402806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,768,0.042641778786977134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,512,0.011239999698268043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,16384,0.030771556827757094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,65536,0.20561956034766304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,12288,0.03080000148879157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,4096,0.007866666548781926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,12288,0.03762400150299072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,128,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,6144,0.10771466626061334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,10240,0.03460888730155097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,128,0.007882666256692674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,128,0.007601777712504069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,256,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,65536,0.06669421990712483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,5120,0.00997155573632982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,7168,0.12448533376057942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,128,0.003957333250178231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,4096,0.5494666629367405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,4096,0.014653333359294467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,2560,0.009296889106432596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,3072,0.030685332086351182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,6144,0.010280888941552904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,6144,0.16128977139790854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,768,0.0057911111248864075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,2560,0.019721777902709115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,2560,0.008298666940795051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,64,0.010249777800507015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,2048,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,16384,0.09380977683597141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,2560,0.05315022336112129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,256,0.0037662221325768363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,2560,0.014448000325096978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,128,0.004106666478845808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,51200,1.1599786546495225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,3072,0.09334577454460992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,3584,0.009937778115272522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,1024,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,1536,0.011050666371981302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,51200,0.05453155438105265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,10240,0.1903902159796821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,64,0.007753777835104201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,16384,0.3640062279171414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,7168,0.23156534300910103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,12288,3.329256057739258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,51200,0.8409982257419162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,6144,0.01844888925552368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,3072,0.008242666721343994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,2560,0.005580444302823808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,128,0.002756444530354606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,51200,0.8321048948499891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,2048,0.01609599921438429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,4096,0.07251022259394328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,256,0.05548000004556444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,16384,0.20136000050438774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,1536,0.07296622461742826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,12288,0.2761182255215115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,6144,0.038768000072903104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,32,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,5120,0.1295484436882867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,768,0.02270488937695821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,7168,0.2885679933759901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,65536,0.4478835529751248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,3072,0.009345778160625035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,512,0.0032808888289663526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,2560,0.018915555543369718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,3072,0.12280089325375027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,3072,0.018256000346607633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,32,0.00267999991774559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,32,0.0026337777574857077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,3072,0.009174221919642555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,10240,0.044996443721983165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,1536,0.052391111850738525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,6144,0.2477048767937554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,65536,1.6830257839626734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,64,0.005886222339338726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,16384,0.04800533254941305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,5120,0.011432888607184092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,2048,0.013077333569526672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,1536,0.01883555617597368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,3584,0.01626044511795044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,16384,22.4590081108941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,6144,0.056186669402652316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,10240,0.18478933970133463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,8192,0.1005857785542806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,1536,0.055276443560918175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,16384,0.03881599836879306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,8192,0.6092302004496256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,12288,0.04759111007054647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,3072,0.006339555399285422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,768,0.008902221918106079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,128,0.01774488886197408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,7168,0.07024088833067152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,3072,0.15674932797749838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,128,0.01687733332316081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,64,0.008980444735950893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,1024,0.0047457776963710785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,51200,0.15212889512379965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,12288,0.050907555553648204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,6144,0.01261333293384976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,128,0.005192000004980299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,16384,0.008105778031879002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,768,0.016113777955373127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,8192,0.06418577829996745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,12288,0.033969776497946844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,6144,0.1478453344768948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,256,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,6144,0.014831999937693277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,1536,0.02069244450993008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,128,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,512,0.004000888930426704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,2560,0.008403555386596257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,65536,0.09697955846786499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,128,0.003067555526892344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,3584,0.08910133441289265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,8192,0.020630222227838304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,7168,0.016329778565300834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,768,0.004521777646409141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,3584,0.5943244298299154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,4096,0.03923111160596212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,51200,0.06202310985989041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,64,0.021203555994563635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,3584,0.21690132882859972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,16384,0.07414488659964667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,1024,0.015253333581818474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,1536,0.010971555279360877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,12288,0.39211376508076984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,3072,0.012421333127551608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,1024,0.005440000030729506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,8192,0.028235554695129395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,3072,0.43885956870185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,1536,0.012527999778588613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,64,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,32,0.004573333180612988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,8192,0.014719999498791166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,2048,0.05858577622307671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,2048,0.07017688618765937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,1536,0.09833066993289524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,32,0.0026275554878844153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,10240,0.05699288845062256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,512,0.018279999494552612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,512,0.0052577778697013855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,64,0.005497777627574072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,3584,0.14741244581010607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,768,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,5120,0.017864000466134813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,4096,0.028520888752407495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,8192,0.026699556244744196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,256,0.0029599999801980127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,1536,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,32,0.021012443635198805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,6144,0.021409778131379023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,12288,8.113811916775173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,2048,0.18560889032151964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,128,0.007086222370465596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,2560,0.028839998775058325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,8192,0.14228355884552002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,5120,0.04209422071774801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,2560,0.006680000159475539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,1024,0.003992889076471329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,8192,0.01401511165830824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,64,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,128,0.014273777604103088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,6144,0.009533333281675974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,7168,0.010604444477293225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,32,0.0029324444217814338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,7168,0.01164533363448249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,51200,0.100545777214898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,4096,0.13996533552805582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,2048,0.019681778219011094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,12288,0.14264444510142008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,12288,0.03253777821858724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,768,0.0058106668293476105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,1536,0.004634666774008009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,5120,0.13413777616288927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,256,0.00444177786509196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,3072,0.01685066686736213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,3584,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,51200,0.11584533585442437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,1536,0.017851556340853374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,3072,0.009893333746327294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,1536,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,6144,0.09158755673302545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,1536,0.1017342209815979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,2048,0.06484622425503202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,512,0.01441333360142178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,128,0.0140346669488483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,5120,0.14243466324276394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,16384,0.3483102321624756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,4096,0.3143155574798584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,2560,0.02570844524436527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,16384,0.06841688685946994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,6144,0.03477599885728624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,3072,0.010110222631030613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,6144,0.006767999794748094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,6144,0.09364000293943618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,128,0.004648888690604104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,65536,0.045385777950286865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,8192,0.09626399808459812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,4096,0.12302754984961616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,5120,0.36048089133368594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,12288,0.007684444387753804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,1536,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,3072,0.10319555468029445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,3072,0.02306933369901445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,65536,2.6633919609917536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,65536,0.12038311693403457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,16384,0.03637866510285272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,8192,1.0192915598551433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,4096,0.40978489981757266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,51200,0.4748071034749349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,3584,0.05513866742451986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,1536,0.007524444825119442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,12288,0.03228266702757941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,4096,0.008936888641781276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,10240,0.01144266708029641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,7168,0.01718311177359687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,8192,0.37629156642489964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,65536,0.28892623053656685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,12288,0.033519112401538424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,768,0.007442666424645319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,64,0.00757422215408749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,512,0.2508329020606147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,768,0.010038221875826517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,4096,0.014300444059901767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,65536,0.012416000167528788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,128,0.030260443687438965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,1024,0.05805955330530802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,3072,0.04624533322122362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,64,0.00300799993177255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,1024,0.1005128886964586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,16384,0.01982399986849891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,65536,3.474641799926758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,5120,0.9231884214613172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,51200,0.09729155566957261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,256,0.01575644479857551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,2560,0.008205333517657386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,51200,1.086073769463433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,64,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,16384,0.050605333513683744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,3072,0.017708443933063082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,64,0.0034924443397257063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,64,0.04701155424118042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,65536,0.5388097763061523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,6144,0.005912888795137405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,16384,0.381402677959866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,128,0.03292622168858846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,7168,0.12812444898817274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,2048,0.013032889200581444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,1024,0.02071555621094174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,51200,0.26547733942667645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,3584,0.03416533271471659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,64,0.008007110820876228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,16384,0.05624177720811632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,51200,0.715261353386773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,1536,0.03168355425198873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,8192,0.023179555932680767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,768,0.004026666697528628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,51200,0.03530044356981913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,4096,0.016207999653286405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,5120,0.459229310353597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,65536,3.476737764146593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,8192,0.04078400135040283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,768,0.04165688819355435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,512,0.008272888759771982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,1536,0.008168000313970778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,32,0.002579555536309878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,256,0.6647849082946777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,16384,0.156058669090271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,8192,0.011720889144473605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,16384,0.02217955556180742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,51200,0.056800888644324414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,128,0.00433688900536961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,6144,0.06487377484639485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,6144,0.30327733357747394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,3584,0.009679111341635386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,5120,0.011167111496130625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,65536,6.044716305202908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,512,0.010031110710567897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,3584,0.02216000027126736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,65536,0.5487422413296169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,1024,0.009831111464235518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,6144,0.021105777886178758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,51200,3.2513056861029734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,12288,0.11911111407809788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,1024,0.00565155554148886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,65536,0.0748986668056912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,16384,0.07479911380343966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,2048,0.0062773335311147906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,768,0.1391351090537177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,2560,0.11111911137898763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,3584,0.02917155623435974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,768,0.004982222285535601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,64,0.007607999775144789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,4096,0.015881778465376962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,5120,0.0136808885468377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,7168,0.015068444940778943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,128,0.004655111167165968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,768,0.045046223534478076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,5120,0.02789688772625393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,6144,0.012728888955381183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,5120,0.006931555353932911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,768,0.055678221914503306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,4096,0.22936712370978463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,3584,0.008971555365456475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,10240,0.15681866804758707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,64,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,256,0.013199999928474426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,128,0.007298666569921706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,7168,0.09265600310431586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,1024,0.005686222265164058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,1024,0.010928000013033548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,2560,0.007083555890454187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,6144,0.032051556640201144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,16384,0.857119984096951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,512,0.012816000315878125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,2560,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,3072,0.23404977056715223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,7168,0.013166222307417126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,256,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,65536,0.09398844507005479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,256,0.018437332577175565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,256,0.028395556741290625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,51200,1.6601351632012262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,2048,0.009820444716347588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,7168,0.03801511062516107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,512,0.027273777458402846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,2560,0.008210666477680206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,32,0.005051555732885997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,3584,0.009465778039561378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,32,0.0031946665710873077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,64,0.0026933333526055017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,16384,0.11859377225240071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,12288,0.7124373118082682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,65536,1.1735795338948567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,128,0.038309332397249006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,128,0.006341333190600078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,2048,0.009071110851234859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,1024,0.014942222171359591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,4096,0.017703110973040264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,6144,0.014502222339312235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,64,0.05388977792527941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,12288,0.018797333041826885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,256,0.01260088880856832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,2048,0.03390844331847297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,256,0.0037573331760035623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,7168,0.01056177748574151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,8192,0.08705155717002021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,7168,0.2777911027272542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,2048,0.11277066336737739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,8192,0.0781351129213969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,16384,0.4402480125427246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,8192,0.04135377870665656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,2560,0.00958577791849772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,16384,0.02973066767056783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,256,0.009201777478059133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,8192,0.06673777765697904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,7168,0.03013333347108629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,1536,0.008424888882372115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,128,0.002943111169669363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,12288,0.4104728963640001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,3584,0.013031110995345645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,1024,0.2195582257376777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,65536,0.10622933175828721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,6144,0.026573333475324843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,51200,0.8323191006978353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,2560,0.00906666699382994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,5120,0.030736890104081895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,128,0.02342222299840715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,3584,0.013153778182135688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,64,0.002578666640652551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,5120,0.20123467180464003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,3584,0.019724445210562814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,4096,0.008914666871229807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,10240,0.01757244434621599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,8192,0.050316443045934044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,2560,0.010616888602574667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,16384,0.03017066584693061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,2048,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,512,0.005223999834722943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,16384,0.013921777407328287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,64,0.004862222406599256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,8192,0.02134399943881565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,32,0.012657778130637275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,6144,0.011571555501884885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,128,0.047482666042115956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,1536,0.027934223413467407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,2560,0.01240888900227017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,12288,0.1541973352432251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,4096,0.10333510902192856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,512,0.008202667037645975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,10240,0.026554667287402686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,6144,0.37824177742004395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,8192,0.11042844586902195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,10240,0.024681778417693243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,256,0.005470222069157495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,2048,0.010596444209416708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,8192,0.02494488822089301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,3584,0.00775822252035141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,128,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,51200,1.5317243999905055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,128,0.0038364442686239877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,16384,0.29487821790907115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,65536,0.07965777979956733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,5120,0.026319111386934917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,128,0.014524444937705994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,51200,0.05715822511249118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,2560,0.00814844419558843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,256,0.0047084444926844705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,256,0.004587555511130227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,2048,0.015586665934986539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,12288,0.039356443617079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,10240,0.4350284470452203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,2048,0.008366222182909647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,1024,0.006281777802440856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,1024,0.08050844404432508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,16384,0.09760622183481853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,2048,0.6689164373609754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,65536,0.262264887491862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,768,0.00674044465025266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,768,0.16284178362952337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,6144,0.013691555294725629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,768,0.0722257759835985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,51200,10.822291056315104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,64,0.009009777671760982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,65536,1.9826880560980902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,65536,4.90160157945421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,5120,0.02831733226776123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,3584,0.008475555313958062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,128,0.006045333213276333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,64,0.00462755560874939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,64,0.041598220666249595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,2048,0.005872000008821487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,2560,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,16384,2.4869537353515625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,4096,0.5158898035685221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,768,0.005623111294375525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,12288,0.7363013161553277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,768,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,16384,0.021159110797776118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,768,0.00982844415638182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,3584,0.00906755526860555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,4096,0.05229333374235365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,7168,0.012472888661755456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,32,0.0027342221389214196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,64,0.005151999907361137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,256,0.011462222370836469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,8192,0.02251200046804216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,64,0.0029120000286234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,5120,0.045440001620186694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,768,0.773412439558241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,10240,0.018153778380817838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,16384,0.03526577684614394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,1536,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,51200,0.4355288876427545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,768,0.009645333720578088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,3584,0.2939564387003581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,2048,0.015487111277050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,51200,0.32763555314805776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,2048,0.46402576234605575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,16384,0.1816791163550483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,7168,0.02242311173015171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,2560,0.04642311069700453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,3584,0.061662223603990346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,2560,0.009158222211731805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,768,0.03361777795685662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,16384,0.05750222338570488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,128,0.1618408891889784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,8192,0.52547820409139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,2048,0.012825777961148156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,64,0.005130666825506422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,2048,0.008216000265545314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,8192,0.2604915565914578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,3584,0.01016622202263938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,3072,0.01575199928548601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,6144,0.07486577828725179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,16384,0.03595022360483805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,10240,0.07221866978539361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,1536,0.006119110931952794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,10240,0.11176978217230903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,2048,0.005391110976537068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,256,0.003690666622585721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,51200,3.4844818115234375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,12288,0.051592889759275645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,12288,0.03250311149491204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,8192,0.3468773365020752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,3072,0.04316622349951002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,6144,1.0741973453097873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,5120,0.18310399850209555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,2560,0.007195555501514011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,2048,0.06762578090031941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,3072,0.1595733298195733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,4096,0.06002755297554863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,65536,0.11673688888549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,512,0.019638222124841478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,8192,0.23063466284010145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,8192,0.31471021970113117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,16384,0.06529244449403551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,65536,3.4574381510416665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,1024,0.01945955554644267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,2560,0.04179288943608602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,3072,0.013653332988421122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,256,0.005664000080691443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,5120,0.13344532913631862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,5120,0.007982222570313348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,12288,6.884915669759114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,12288,0.11603555414411758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,512,0.01573866605758667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,7168,0.5473848978678385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,1536,0.0424924459722307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,1024,0.14952445030212402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,65536,0.07570933633380465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,4096,0.4555688963996039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,7168,0.8674880133734809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,3584,0.011191111471917896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,64,0.003640000190999773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,5120,0.11704711119333903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,3072,0.011338666909270816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,4096,0.018629334039158292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,4096,0.015669332610236276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,65536,0.09897244638866848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,3584,0.18291021717919242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,6144,0.01460977726512485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,7168,0.013275555438465543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,12288,0.018603555030292936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,32,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,3584,0.024680889315075342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,1536,0.030440890126758154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,256,0.004395555704832077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,128,0.015184000134468079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,3072,0.02077866593996684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,32,0.008943110704421997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,10240,0.02440533373090956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,3584,0.3232213391198052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,2048,0.007521777517265744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,2048,0.01365600029627482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,5120,0.06921866867277357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,3072,0.0849226646953159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,3072,0.013927999469969006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,512,0.016972444123691983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,64,0.0026488889836602737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,2048,0.0070879997478591064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,1024,0.004231111043029361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,2560,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,16384,0.02179733415444692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,4096,0.4130906793806288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,6144,0.44717688030666775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,12288,0.018387556076049805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,32,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,4096,0.011298666397730509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,7168,0.17575644122229683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,10240,0.0240346673462126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,7168,0.030080888006422255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,768,0.007697777615653143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,7168,0.19180356131659615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,3072,0.4480213589138455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,12288,0.07895733250512017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,65536,6.57003190782335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,7168,0.021176000436147053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,768,0.007154666715198093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,16384,0.11581510967678493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,1536,0.011271110839313932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,10240,0.043988443083233304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,12288,0.0894426637225681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,7168,0.11435466342502171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,3072,0.02699733277161916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,2048,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,1024,0.04202222161822849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,6144,0.027288888891537983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,65536,0.07934221956464979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,16384,0.6075421969095866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,5120,0.0473653342988756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,8192,0.013869333598348828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,10240,0.009847111172146266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,16384,0.04190133346451653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,1536,0.01828977796766493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,128,0.0069297779765393995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,32,0.018132444885041978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,65536,6.278746710883246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,4096,0.008952000074916417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,512,0.007887110941939885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,1536,0.04179733329349094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,10240,0.04599555664592319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,32,0.005144888742102517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,7168,0.1384177737765842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,7168,0.14124800099266901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,3584,0.007896000312434303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,768,0.005895111295912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,3584,0.06893510950936212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,51200,0.08483288685480754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,12288,0.54603820376926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,2048,0.061815996964772545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,10240,0.17244177394443086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,2560,0.006823110911581252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,3072,0.03327911098798116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,6144,0.012279111478063794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,1536,0.056463109122382275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,10240,0.013378666506873237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,51200,1.6373120413886175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,51200,0.07467289103402032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,7168,2.25492180718316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,2560,0.06879555516772799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,6144,0.011584888729784222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,7168,0.03596977723969354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,51200,9.792801751030817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,1536,0.013761777844693927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,128,0.004041777716742622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,65536,0.1615742180082533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,65536,3.3175830841064453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,32,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,128,0.003301333429084884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,6144,0.017902221944597032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,8192,0.14620533254411486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,51200,0.47214757071601016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,65536,0.30892711215549046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,65536,0.8184844652811686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,10240,0.02402755618095398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,256,0.04866577850447761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,3072,0.010933332973056369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,7168,0.31519556045532227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,10240,0.017477333545684814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,2048,0.008583111067612966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,7168,0.010614222122563256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,3072,0.15379110972086588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,51200,0.054195556375715465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,7168,0.027614222632514104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,5120,0.010022222167915767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,10240,0.03434044453832839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,1024,0.011868444581826528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,1024,0.0042239998777707415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,3584,0.028620445066028174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,65536,0.07470844189325969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,512,0.00850311087237464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,32,0.005381333331267039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,2048,0.041570666763517596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,128,0.00423200014564726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,5120,0.012798222402731577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,768,0.06614310873879327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,128,0.014662222729788886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,6144,0.00977422214216656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,1536,0.007666666474607255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,7168,0.012518222133318583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,6144,0.03109955456521776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,64,0.007377777662542131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,1024,0.10452533430523342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,4096,0.07988533046510485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,1024,0.05268889003329807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,12288,0.043375111288494535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,128,0.004589333302444882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,8192,1.4923715591430664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,1536,0.017431111799346078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,128,0.009566222627957663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,1536,0.008000888758235509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,10240,0.01993600030740102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,3584,0.011584888729784222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,1536,0.07624710930718316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,8192,0.11172444290584987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,768,0.02215555475817786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,3072,0.01127199994193183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,128,0.003286222202910317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,8192,0.0893733302752177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,16384,0.4541235499911838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,16384,0.05227466755443149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,3072,0.022876444790098403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,64,0.0284862220287323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,4096,0.011493333511882357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,1024,0.010149333212110732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,64,0.10889599720637004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,512,0.011624000138706632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,512,0.29819199774000377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,65536,0.0736337767706977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,6144,0.011702222128709158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,512,0.00738844441043006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,3072,0.04941777719391716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,3584,0.00870400004916721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,6144,0.16187911563449436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,2560,0.013099555340078143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,16384,1.7346969180636937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,2560,0.04630400074852837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,7168,0.06879378027386136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,1536,0.006135111053784688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,256,0.005399111244413588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,5120,0.010406222608354356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,3584,0.12656177414788142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,65536,0.5543680191040039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,5120,0.014888000157144336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,32,0.008395555946562026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,256,0.01165244479974111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,3584,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,51200,0.0477786660194397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,5120,0.37453601095411515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,512,0.005555555638339784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,51200,0.01313777764638265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,1024,0.024059555596775476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,3072,0.03283199999067519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,8192,0.17248088783688018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,12288,0.653547551896837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,51200,14.808302137586805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,768,0.18781777222951254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,256,0.014956444501876831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,6144,0.09896800253126357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,3584,0.00776533368561003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,4096,0.13589333163367376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,16384,0.012801777985360889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,65536,7.479617648654514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,2048,0.6925990846421984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,3072,0.033936887979507446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,6144,0.03759555684195624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,32,0.006815111057625876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,4096,0.011286222272449069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,256,0.005154666801293691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,512,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,51200,0.4974524180094401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,64,0.003050666716363695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,64,0.047019554509056934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,4096,0.1382746696472168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,10240,0.029001779026455347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,12288,0.8804817729526095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,1536,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,65536,0.07656621932983398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,3072,0.02145066691769494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,768,0.0123857781291008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,4096,0.10895199908150567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,51200,0.06915466652976142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,128,0.01867466668287913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,10240,0.015654222832785714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,6144,0.1264222198062473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,7168,2.828616036309136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,8192,0.02679466704527537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,2048,0.6468533409966363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,3072,0.022336888644430373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,2560,0.05425688955518934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,2560,0.02086755633354187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,512,0.022887999812761944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,2048,0.005351999981535806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,6144,0.009081777599122789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,2048,0.008475555313958062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,128,0.002720000015364753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,3584,0.051049778858820595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,1024,0.2806933455997043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,3072,0.13130577405293783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,2560,0.041272888580958046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,2048,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,32,0.005657777604129579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,1536,0.07214666737450494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,3584,0.007920888562997183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,7168,0.1826124456193712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,3584,0.053820444477929004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,7168,0.031054221921496924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,3072,0.009948444035318162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,256,0.0057706667317284485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,10240,0.10988089111116196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,4096,0.16632978121439615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,8192,0.0069057775868309875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,65536,0.3857617908053928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,512,0.21340533097585043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,64,0.0036799998746977914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,7168,0.8594702084859213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,4096,0.05305955476231045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,512,0.025432889660199482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,256,0.013279111021094851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,6144,0.10255999697579278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,3072,0.011906666888131036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,3072,0.020308444897333782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,128,0.008220444122950235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,1024,0.005672889037264719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,1024,0.008348444269763099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,16384,0.01851377718978458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,8192,0.009689778089523315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,2048,0.009375111096435124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,32,0.0052177777720822235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,5120,0.510058667924669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,768,0.02033333314789666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,5120,0.040770666466818914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,2560,0.04174133472972446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,2048,0.04600355691379971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,65536,4.613161722819011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,1024,0.006755555669466655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,10240,0.012683555483818054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,7168,0.022192888789706763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,4096,0.023221333821614582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,12288,0.016953777935769822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,3584,0.8507795333862305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,51200,0.5007022221883138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,64,0.014113777213626437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,10240,0.016062221593326993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,16384,0.0647875534163581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,768,0.005211555709441503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,128,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,51200,0.1360728873146905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,4096,0.07854755719502766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,1024,0.004015999949640698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,12288,0.03820888863669501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,64,0.009956444303194681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,6144,0.013593778014183044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,2048,0.008220444122950235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,5120,0.015597333510716757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,32,0.01960355540116628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,64,0.008553778131802877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,512,0.011644444531864591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,4096,0.02202488813135359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,768,0.01705511079894172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,512,0.07881422175301446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,8192,1.1241608725653756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,6144,0.9760239919026693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,7168,0.016864000095261466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,12288,0.032623112201690674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,3072,0.08400355445014106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,3072,0.04321066538492838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,2560,0.01295199990272522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,768,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,51200,1.2030870649549696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,128,0.007659555309348636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,12288,0.13298132684495714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,10240,0.2550133334265815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,6144,0.07463288969463773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,1024,0.004839999808205498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,8192,0.3757164478302002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,1536,0.040023999081717596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,2560,0.009944889280531142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,51200,26.218339708116318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,2560,1.5077084435356989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,128,0.008948444492287105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,512,0.008747555315494537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,512,0.004776000148720211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,2048,0.015180443723996481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,3584,0.025182222326596577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,16384,5.230726030137804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,1024,0.012055111428101858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,65536,0.9153928756713867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,4096,0.01550044450494978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,128,0.002970666728085942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,6144,0.3331564532385932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,2048,0.03916800022125244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,256,0.0057555557125144536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,3072,0.03828088773621453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,6144,0.008873778084913889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,5120,0.016102222932709586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,512,0.006096000058783426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,6144,0.020423110988405015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,2048,0.5556071069505479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,768,0.10645333263609146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,7168,0.015541333291265698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,2560,0.011218667030334473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,6144,0.019685332973798115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,6144,0.37851643562316895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,1024,0.013071111506885953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,1536,0.011953777737087674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,128,0.007725333174069722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,512,0.0053680001033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,2560,0.01590755581855774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,5120,0.009300444689061906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,5120,0.3308488792843289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,3072,0.3425244490305583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,768,0.0037413334680928122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,64,0.0028604444944196274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,3072,0.007981333467695449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,1536,0.0069546666410234236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,256,0.005426666802830166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,12288,4.950652652316623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,12288,0.8285937839084201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,64,0.003703111161788305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,8192,0.013944000005722046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,3584,0.02179199953873952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,4096,0.04668710960282219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,51200,0.21982933415306938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,16384,0.05782044596142239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,128,0.0037635556525654266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,7168,0.03378222386042277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,12288,1.7070568932427301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,65536,0.09741510947545369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,5120,0.033506668276256986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,8192,0.0972248911857605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,7168,0.027971555789311726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,51200,0.3167475594414605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,4096,0.020847110284699336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,2048,0.09002133210500081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,51200,1.6862337324354384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,65536,1.8189154730902777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,6144,1.7889618343777125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,10240,0.02240444388654497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,3584,0.008540444076061249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,12288,0.4632906913757324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,3584,0.10815911160575019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,5120,0.0069306666652361555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,10240,0.5213671260409886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,256,0.05221244361665514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,8192,0.12790666686164007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,256,0.19209244516160753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,3072,0.016125332978036668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,64,0.0032702222880389956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,768,0.005976000179847081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,512,0.00609777785009808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,128,0.0035146665241983202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,10240,0.9302071465386285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,5120,0.027533332506815594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,2048,0.009639999932712978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,51200,0.41726043489244247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,16384,0.09806311130523682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,3072,2.529038323296441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,6144,0.1555280023150974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,1536,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,4096,1.2412506739298503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,12288,0.28727467854817706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,1536,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,12288,0.02572622232966953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,16384,1.2911769019232857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,7168,0.26616978645324707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,6144,0.18572710620032418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,4096,0.2045288880666097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,51200,0.04535911149448819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,8192,0.033518221643235944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,32,0.005757333503829108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,51200,0.06249511241912842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,51200,0.9228880140516492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,16384,0.2088142236073812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,6144,0.32841510242886013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,2560,0.021996445126003687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,3584,0.08406044377221002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,2048,0.010096000300513374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,7168,0.052226665947172374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,1024,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,768,0.05720533265007866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,7168,0.013004444539546967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,32,0.0027511111564106415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,32,0.003136000078585413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,12288,0.12430044015248616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,65536,0.06804800033569336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,2560,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,6144,0.09868088695738052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,3584,0.07986311117808025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,10240,0.31422045495775014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,7168,3.497489717271593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,1536,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,65536,0.07438311311933729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,65536,5.114226871066623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,7168,0.028612444798151653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,1536,0.0352497763103909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,32,0.0027751111321979095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,7168,0.020569778150982328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,64,0.0031431110368834603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,6144,0.029001779026455347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,16384,0.045147554741965405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,768,0.0052044445441828836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,2048,0.011691555380821228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,3072,0.018741332822375827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,1024,0.013696889082590738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,7168,0.017311111092567444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,256,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,12288,0.022423999177085027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,7168,0.0661733349164327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,64,0.009434666898515489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,8192,0.05948978000217014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,10240,0.016160888804329764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,768,0.4469555748833551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,5120,0.2516355514526367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,16384,0.3860373232099745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,3072,0.0955431130197313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,6144,0.02201777696609497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,512,0.009224889179070791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,768,0.008942222429646386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,1024,0.10532089074452718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,16384,0.010690666735172272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,768,0.009333333207501305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,64,0.002609777781698439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,512,0.1143173376719157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,12288,0.09277511305279201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,3072,0.016207110550668504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,256,0.003072888900836309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,4096,0.014427555931939019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,16384,0.019645333290100098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,512,0.10882933272255792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,65536,1.0795155631171334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,3072,0.024804444776640996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,4096,0.014750222365061441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,51200,0.09570488664839004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,2048,0.14794845051235622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,5120,0.015257777439223396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,1024,0.05048266715473599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,10240,0.015023110641373528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,64,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,7168,0.017954667409261067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,7168,0.11752533912658691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,8192,0.141229338116116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,2048,0.22780267397562662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,8192,0.012410666379663678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,768,0.0038008888562520346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,3584,0.052469333012898765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,16384,0.05954044395022922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,256,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,32,0.0041991112132867174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,6144,0.013183111117945777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,1024,0.017161778277821012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,3584,0.1519813272688124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,3584,0.023749333288934495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,8192,0.014257777068350049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,2560,0.01571466690964169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,4096,0.03555555476082696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,1024,0.02712000078625149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,7168,0.2011786699295044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,768,0.012890666723251343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,32,0.02485688858562046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,3072,0.012649777862760754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,4096,0.010409778190983666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,12288,0.4001679950290256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,1536,0.12765333387586805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,4096,0.041195554865731135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,5120,0.02511288887924618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,32,0.018851555056042142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,2560,0.022240888741281297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,128,0.17391822073194715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,256,0.008910222185982598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,5120,0.011839110818174152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,65536,0.09848533074061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,51200,0.17711910936567518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,2048,0.017528888252046373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,2560,0.008604444563388824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,65536,0.657184018029107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,512,0.004872888740566042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,128,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,64,0.027179555760489568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,6144,0.019526221685939364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,51200,0.6092115508185493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,4096,0.16129867235819498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,768,0.009182222187519073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,4096,0.008744888835483128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,64,0.0027662221756246355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,1024,0.016141333513789706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,2560,0.011626666618718041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,5120,0.08001333475112915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,7168,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,10240,0.014025777578353882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,7168,0.17227644390530059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,12288,0.05017955435646904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,1024,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,1536,0.19244799349043104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,65536,0.37465688917371964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,7168,0.031113776895735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,256,0.0033475555893447665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,3072,0.13946933216518825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,16384,0.434371550877889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,65536,8.302313062879774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,12288,0.5333671040005153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,128,0.0053084443012873335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,4096,0.13548266887664795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,32,0.016952888833151925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,1536,0.11666844950781928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,768,0.004917333523432414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,2048,0.4836151334974501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,2560,0.035410665803485446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,6144,0.01605422298113505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,768,0.00378222225440873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,128,0.010100444157918295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,7168,0.027526222997241553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,10240,0.06565066840913561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,2048,0.12479821840922038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,512,0.0033253334048721525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,65536,0.24436622195773652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,8192,0.011547555526097616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,64,0.011546666423479715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,3072,0.018039110634062026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,8192,0.015421332584487068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,128,0.006002666635645761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,6144,0.03567377726236979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,5120,0.011407111254003314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,32,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,768,0.003912000192536248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,51200,0.23141956329345703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,2560,0.6282826529608833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,51200,3.8391456604003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,8192,0.018774222996499803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,10240,0.006776888751321369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,6144,0.7187831136915418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,512,0.00611200018061532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,16384,0.06235822041829427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,768,0.00574222207069397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,65536,0.1103973322444492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,1024,0.024480889240900677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,1024,0.010746666954623329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,32,0.005596444424655702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,12288,0.2456080118815104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,64,0.003346666693687439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,51200,1.332367155287001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,8192,0.019091554813914828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,256,0.05108177661895752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,3072,0.2102888955010308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,51200,0.20030311743418375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,6144,0.7117066913180881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,3584,0.11799821588728164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,3584,0.01607022186120351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,4096,0.04702933298216926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,768,0.00775733341773351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,4096,0.014663999279340109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,128,0.0047146665553251905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,2560,0.010079999764760336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,12288,0.29640089140997994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,3072,0.3084008958604601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,64,0.008597333398130205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,1024,0.00583022211988767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,12288,0.017599110802014668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,256,0.16453066137101915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,2048,0.00814933329820633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,1536,0.004815111143721474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,256,0.013960000541475085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,3584,0.02470933397610982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,512,0.00443466669983334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,12288,0.20651822619967988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,2560,0.012004444168673621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,3584,0.008790222307046255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,8192,0.2342017756568061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,2048,0.007175111108356052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,3072,0.022602667411168415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,512,0.003277333246337043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,6144,0.014597333139843412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,8192,0.01705422169632382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,10240,0.029579556650585596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,3072,0.008618666893906063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,2560,0.1236906713909573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,10240,0.08202399810155232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,64,0.005969777703285217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,6144,0.007082666787836287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,4096,0.009883555273214975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,2560,0.032476445039113365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,1536,0.007870222131411234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,256,0.0221368885702557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,7168,0.09930577543046738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,16384,0.9420329199896919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,8192,0.008370666868156856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,768,0.010518222219414182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,5120,1.8056382073296442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,2048,0.007224889265166388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,3072,0.014313777287801107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,51200,0.03355288836691115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,8192,0.009200000100665623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,12288,1.574455155266656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,1536,0.033486220571729876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,65536,8.022921244303385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,12288,1.1797653834025066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,2048,0.0765297810236613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,6144,0.4023768901824951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,51200,22.902976142035588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,256,0.004375111311674118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,1024,0.004171555654870139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,8192,0.07421866390440199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,64,0.002567111204067866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,2560,0.11065955294503106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,51200,3.6196746826171875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,3072,0.019268444842762418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,512,0.023387556274731953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,7168,2.5573529137505426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,10240,0.5558764669630263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,1024,0.007074666519959767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,6144,0.035679999325010515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,3072,0.009922666682137383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,6144,0.028355555401908025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,2560,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,5120,0.44730223549736875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,2560,0.017283555534150865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,3072,0.05840355820126004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,3584,0.016159999701711867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,10240,0.0069919998447100324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,256,0.004368888835112254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,10240,0.1801555554072062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,65536,0.07814400063620673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,512,0.0063075555695427795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,2560,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,1536,0.0326533317565918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,512,0.003998222036494149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,64,0.007072889142566257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,128,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,65536,0.21059377988179526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,2560,0.017716444200939603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,3584,0.1270151138305664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,1024,0.014728888869285583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,3072,0.010574222438865237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,2560,1.1794746186998155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,32,0.011594666375054253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,7168,0.12545777691735163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,3584,0.007618666523032718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,3584,0.0986479984389411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,12288,0.2730933295355903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,6144,0.01912088857756721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,4096,0.00941244430012173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,2560,0.01902577777703603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,768,0.0037413334680928122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,51200,1.7627368503146703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,1536,0.05912711222966512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,2560,0.02444444431198968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,7168,0.1781031158235338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,3072,0.011710222396585675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,10240,0.026556443836953905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,2560,0.014583110809326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,256,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,7168,0.288445340262519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,1536,0.009191999832789103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,32,0.006067555397748947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,512,0.005655111124118169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,64,0.0978871120346917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,12288,0.046698666281170316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,7168,0.007542221910423703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,3072,0.021186666356192693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,65536,0.24558043479919434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,4096,0.06163733535342746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,2048,0.016303999556435477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,5120,0.04979733294910855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,16384,8.967701382107204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,3072,0.008630221916569604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,12288,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,1024,0.05788266658782959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,65536,1.5698853598700628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,768,0.01685599982738495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,12288,0.03631110986073812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,64,0.035700443718168474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,256,0.07795111338297527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,3584,0.00998222248421775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,7168,2.8982870313856335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,128,0.03030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,64,0.02245866590076023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,512,0.016568000117937725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,8192,0.14275377326541477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,6144,0.01682222220632765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,1536,0.007401777638329401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,12288,0.017459554804695975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,256,0.007663999994595845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,2048,0.013773333695199756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,768,0.06064266628689236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,128,0.004512888689835866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,1536,0.005234666582610872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,65536,0.5243093172709147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,16384,0.03921866748068068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,4096,0.1534693373574151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,5120,0.0693439973725213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,12288,0.09323377741707696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,12288,0.033583111233181424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,4096,0.037138667371537946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,256,0.023256889647907678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,1536,0.006735999965005451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,2560,0.010844444235165914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,64,0.0029128889242808023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,6144,0.010298666854699453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,5120,0.0184871107339859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,2560,0.012255111502276527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,5120,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,768,0.034472889370388456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,4096,0.07204444540871514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,1024,0.004088888979620404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,3584,0.22402845488654244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,2048,0.012258666257063547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,6144,0.009655111365848118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,3072,0.027273777458402846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,32,0.002583111118939188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,65536,0.06831200255288018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,1536,0.012607999973826937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,6144,0.09085688988367717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,512,0.08202577961815728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,12288,0.6744089126586914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,256,0.0075288886825243635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,1024,0.011593778100278644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,4096,0.021711111068725586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,7168,0.5179991192287868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,16384,0.06096000141567654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,2048,0.013714666995737286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,5120,0.10160266690784031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,2048,0.0075084442893664045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,16384,0.015024888846609326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,64,0.010089777410030365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,2560,0.007438222567240397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,256,0.010298666854699453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,512,0.017927999297777813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,3072,0.13910310798221165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,1536,0.02139288849300808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,65536,0.09478311406241523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,2048,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,512,0.0035208890007601846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,32,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,256,0.005446222093370225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,65536,0.2027191056145562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,3072,0.005986666513813867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,128,0.022631999519136217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,7168,0.5048346519470215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,4096,0.03625599874390496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,51200,0.09659288989173041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,512,0.003524444583389494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,1536,0.0052622221410274506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,1024,0.03670310974121094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,3584,0.0922586652967665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,128,0.011830222275522021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,512,0.005884444548024072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,6144,0.057180444399515785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,12288,0.06879822413126628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,65536,0.17890577846103242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,512,0.018769777483410306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,51200,0.014384888940387301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,4096,0.011249778171380361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,64,0.005152889009979037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,65536,0.33523021803961855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,32,0.0075804442167282104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,65536,0.2816026740603977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,512,0.00555288874440723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,8192,0.06694400310516357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,2048,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,64,0.041542222102483116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,8192,0.18317777580685088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,256,0.04167111052407158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,768,0.01941155559486813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,2048,0.08851022190517849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,16384,0.10990933577219646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,7168,0.013217777841620974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,8192,0.06598399745093451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,3072,0.11321600278218587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,6144,0.06229066848754883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,10240,3.3153262668185763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,768,0.009257777697510189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,4096,0.020872889293564692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,3584,0.2070337798860338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,51200,6.4850586785210504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,3584,0.011165333290894827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,1536,0.05406044589148628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,10240,0.5312168863084581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,64,0.004297777596447202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,2048,0.017656000124083627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,2560,0.007813333637184566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,12288,0.06722133027182685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,768,0.05314844515588549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,3584,0.06386666827731662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,1024,0.006197333335876465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,6144,0.005948444621430502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,5120,0.012135999898115793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,128,0.007297777467303806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,768,0.006275555739800136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,1536,0.4303910997178819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,256,0.010915555887752108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,8192,0.011938666303952536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,7168,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,3072,0.008600888980759515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,2048,0.014058666096793281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,3072,0.06217510832680596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,7168,0.019443555010689628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,32,0.004105777790149053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,12288,0.030878222650951807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,12288,0.04138933287726508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,5120,0.03929333223236932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,2048,0.008355555435021719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,3584,0.08457244767083062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,7168,0.12547466490003797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,12288,0.024668445189793903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,2560,0.020249777370029025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,768,0.06399110953013103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,16384,0.12338933679792617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,8192,0.03189155459403992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,4096,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,2560,0.08344355556699966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,16384,0.20631199412875703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,512,0.0058088890380329555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,65536,0.9081083933512369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,256,0.0038195554580953387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,2560,0.028055999014112685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,8192,0.3156764507293701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,6144,0.05785333447986179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,1536,0.08169155650668673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,1536,0.0047208890318870544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,1536,0.0073520003093613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,1536,0.17203822400834826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,256,0.0037342223028341928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,10240,1.6467644373575847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,128,0.016955556141005624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,1536,0.011121778024567498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,5120,0.019095111224386428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,6144,0.2289422353108724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,32,0.0026559999419583213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,3584,0.010633777413103314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,51200,0.23162844445970324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,65536,0.08858666817347209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,256,0.0053484443989064955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,2560,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,512,0.005678222411208683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,1024,0.011319110790888468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,51200,1.136151949564616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,8192,0.21343821949428984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,1024,0.012118221984969245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,1536,0.036454220612843834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,32,0.04031822085380554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,3584,0.008894222478071848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,512,0.007231999602582719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,3072,0.01856355534659492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,51200,1.028512848748101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,3072,0.08541155523724026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,1024,0.012200000385443369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,16384,0.05366933345794678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,6144,0.5194613138834635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,12288,0.05117422342300415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,1536,0.011206222077210745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,768,0.004345777961942884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,3584,0.01622577839427524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,1024,0.022807111342748005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,8192,0.45397424697875977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,2560,0.007329777710967594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,1024,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,8192,0.020011555817392137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,8192,0.010048000348938836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,2048,0.019303111566437613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,512,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,64,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,1024,0.05411200059784783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,6144,0.020198222663667467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,3584,0.008830221990744272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,12288,0.2571635511186388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,10240,0.028104000621371802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,2048,0.07653511232799955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,3584,0.024539555112520855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,6144,0.02204533252451155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,256,0.034449779325061373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,65536,0.28842488924662274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,51200,0.10664711395899455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,2048,0.09568622377183701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,3072,0.3481608761681451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,128,0.04446755515204536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,128,0.004015111260943943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,1536,0.07156088617112902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,768,0.009834667046864828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,3072,0.009327111144860586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,1536,0.05009422368473477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,1024,0.055330667230818004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,3072,0.26584000057644314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,1536,0.0047280001971456735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,4096,0.15796800454457602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,12288,0.18165955278608534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,768,0.07176266776190864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,6144,0.030991110536787245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,3072,0.021394666698243883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,3072,0.3133653269873725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,16384,0.03933066791958279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,10240,0.02409066590997908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,3584,0.09990044434865315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,5120,0.04853066802024841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,128,0.026849778162108526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,8192,0.2496702141231961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,64,0.002663111107216941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,1024,0.01033688916100396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,7168,0.09998844729529487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,1536,0.05921689007017347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,10240,0.2895840009053548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,1536,0.010603555374675326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,5120,0.02863911125395033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,10240,0.08844444486829971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,10240,0.12246399455600315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,256,0.00462755560874939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,12288,0.01796711153454251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,3584,0.009570666485362584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,256,0.005246222019195557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,6144,0.20054933759901258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,5120,0.07572978072696261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,768,0.0058417779703934984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,768,0.010464888479974536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,10240,0.013479111095269522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,51200,0.05327644281917148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,2560,0.007101332975758447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,2560,0.005624889085690181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,2560,0.014703111516104804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,32,0.00500177757607566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,5120,0.006601777755551868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,768,0.01070755554570092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,51200,0.08455733458201091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,128,0.014295111099878946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,3072,0.051254222790400185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,4096,0.01752800080511305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,64,0.0026977778308921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,7168,0.049639112419552274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,2560,0.014504889647165934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,16384,0.05091999967892965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,5120,0.01924711134698656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,128,0.004915555732117759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,7168,0.08922311332490708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,128,0.004056888735956616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,3072,0.1009662217564053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,1024,0.0410444438457489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,1024,0.014274666706720987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,12288,0.02478933334350586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,7168,0.0464826656712426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,2048,0.07944977945751615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,1536,0.03446933295991685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,256,0.017800889081425138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,4096,0.009863110880057016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,10240,0.01591377788119846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,256,0.0052177777720822235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,10240,0.061624891228146024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,128,0.004936000125275718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,128,0.00416088890698221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,51200,16.896686130099827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,512,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,512,0.024315555890401203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,16384,0.05548710955513848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,64,0.004074666649103165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,2560,0.015223999818166098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,5120,0.008610666626029545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,12288,0.017462222112549674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,7168,0.9756355285644531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,1536,0.012724444270133972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,16384,0.2865973313649495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,256,0.017440888616773818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,2048,0.0074746666683091064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,6144,0.032784889141718544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,4096,0.024495111571417913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,2560,0.007551111280918121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,512,0.006151999864313338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,3584,0.006981333096822103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,32,0.13609600067138672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,16384,1.7331013149685328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,51200,0.12229777706993951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,16384,0.019681778219011094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,256,0.022894221875402663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,7168,4.29234864976671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,5120,0.046895109944873385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,3584,0.0088764445649253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,3584,0.007069333559936947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,256,0.007304000357786815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,3072,0.20528799957699248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,51200,0.05820888943142361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,1536,0.009383111364311641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,8192,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,51200,0.18711377514733207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,1024,0.004398222184843487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,10240,0.01847466660870446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,7168,0.01407911048995124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,65536,0.18038754993014863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,10240,0.02075199948416816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,256,0.010229333407349056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,2560,0.03580177823702494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,65536,0.08363999923070271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,65536,0.20883911185794404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,7168,0.014704000618722705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,3072,1.7627475526597765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,12288,0.26957776811387807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,5120,0.01778133379088508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,2560,0.03744355506367154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,1536,0.10388266377978855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,64,0.00872622181971868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,3584,0.08928800291485256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,1536,0.014280888769361707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,2048,0.005414222263627582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,512,0.008914666871229807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,8192,0.359369781282213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,4096,0.27090666029188365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,3584,0.04303644431961907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,51200,13.680055406358507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,32,0.0042275554604000514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,51200,1.1368684768676758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,51200,0.13792978392706975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,8192,0.04156711035304599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,1024,0.004134222037262387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,768,0.003773333297835456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,8192,0.04618844389915466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,3072,0.025148444705539282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,768,0.004829333474238713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,2560,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,51200,0.01221333361334271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,768,0.010247111320495605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,7168,0.012104888757069906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,2560,0.08474666542477077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,512,0.003452444449067116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,2560,0.014289778139856128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,1536,0.00720355576939053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,32,0.0029279999434947968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,7168,0.010776888993051318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,8192,0.02665155629316966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,1024,0.02779200010829502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,10240,0.016362667083740234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,5120,0.010078222387366826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,2048,0.008548444343937768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,512,0.0038684445122877755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,1024,0.014807111687130399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,7168,0.04063911239306132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,1024,0.022321777211295232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,512,0.012736000120639801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,51200,0.055459555652406484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,16384,0.3661048942142063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,8192,0.02439911166826884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,2560,0.009895111123720804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,768,0.005403555515739653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,32,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,7168,0.05079288946257698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,16384,0.04714666803677877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,2560,1.67920536465115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,51200,0.11826399962107341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,5120,0.007710222568776872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,3072,0.1250391138924493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,8192,0.015314666761292351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,768,0.006233777850866318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,5120,0.012381333443853589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,64,0.004311999926964442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,128,0.00453599997692638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,1024,0.016136000553766888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,2048,0.005447111195988125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,768,0.03283377819591098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,12288,0.017616000440385606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,512,0.01365066650840971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,4096,0.07775733206007215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,7168,0.380528900358412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,1024,0.0046160001721647055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,32,0.005605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,1024,0.044158223602506846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,16384,0.10497244199117024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,3072,0.010135110881593492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,5120,0.013059555656380124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,12288,23.62747361924913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,1024,0.35982844564649796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,4096,0.19669689072502983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,51200,0.26781243748135036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,512,0.010407999985747868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,6144,20.544425116644966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,256,0.008381333616044786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,768,0.0047644442982143825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,5120,0.015541333291265698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,51200,0.9297493828667535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,128,0.00462311092350218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,512,0.0421999990940094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,65536,0.06573333342870076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,12288,0.01924444403913286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,64,0.00510044437315729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,768,0.006343110981914733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,64,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,6144,0.02444177700413598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,2560,0.006186666587988536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,51200,1.1792515648735893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,64,0.0069431112044387394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,5120,0.014087110757827759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,16384,0.060117330816056996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,8192,0.02987288766437107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,51200,1.9562888675265844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,16384,0.08167466852400038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,5120,0.027573333846198186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,32,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,2560,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,6144,1.8360595703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,2048,0.007183999651008182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,512,0.03051911128891839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,12288,0.04203644394874573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,1024,0.006115555349323485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,16384,0.8009084595574273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,10240,0.017772444420390658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,256,0.003953777667548922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,1024,0.040483554204305015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,6144,0.23525955941942003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,7168,0.014009777042600842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,1536,0.00612088872326745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,256,0.004008000095685323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,2560,0.01641244524055057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,4096,0.16525422202216253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,7168,0.10804533296161228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,3584,0.008823999928103553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,32,0.011554666691356234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,7168,0.04671555426385668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,3584,0.30068445205688477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,1536,0.06757155391905043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,512,0.004535111288229625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,10240,0.14740088250901964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,768,0.008605333666006723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,10240,0.04169599877463447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,32,0.01035644445154402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,768,0.014863111906581454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,65536,0.09879644711812337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,16384,0.39108800888061523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,3072,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,3072,0.03393955694304572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,64,0.0037324445115195382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,1536,0.0324248903327518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,8192,0.006514666808976068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,128,0.1471760008070204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,256,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,16384,0.06887733274035983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,6144,0.07613688707351685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,768,0.017321777012613084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,65536,0.5496186680263943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,5120,0.02258666687541538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,8192,0.05462133222156101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,2560,0.0069191112286514705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,8192,0.025817778375413682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,65536,2.2106959025065103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,256,0.034196443027920194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,768,0.009374221993817223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,3072,0.02611911131276025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,3072,0.10432977808846368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,768,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,512,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,2560,0.01757955551147461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,5120,0.013552889227867126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,2560,0.09397688839170669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,256,0.01184088902340995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,51200,1.9845394558376734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,51200,1.6366568671332464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,256,0.004020444634887907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,768,0.017116443978415597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,64,0.006778666542636023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,10240,0.019818666908476088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,512,0.02218666672706604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,128,0.004930666751331753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,3584,0.02351466649108463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,16384,6.521672990587022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,7168,0.23381243811713326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,7168,2.1322996351453996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,4096,0.1626302268770006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,5120,0.007205333146784041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,1024,0.0551111102104187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,768,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,128,0.006248000181383557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,768,0.005759999983840519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,10240,0.7718275388081869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,32,0.014561777313550314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,1536,0.011504000259770287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,16384,0.04970577690336439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,16384,2.294862323337131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,3584,0.009229333036475712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,7168,0.2926959991455078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,51200,0.8224586910671658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,512,0.005094222310516569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,51200,0.060825778378380664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,16384,0.18511288695865205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,32,0.21062400605943468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,32,0.03773955504099528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,12288,1.3335786395602758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,1024,0.025690666503376428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,3072,0.023980443676312763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,1024,0.014866666661368476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,4096,0.04492355717553032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,1024,0.008983111215962304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,256,0.005998222364319696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,256,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,51200,1.6125493579440648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,3584,0.0101742222905159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,64,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,3584,0.3590106699201796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,64,0.10674044158723618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,128,0.004048000193304486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,8192,0.014997333288192749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,2560,0.008479110896587372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,12288,0.3201395670572917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,2048,0.01054044481780794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,65536,0.12941511472066244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,64,0.008103110724025303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,10240,0.08912799755732219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,512,0.02894222074084812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,16384,0.18926933076646593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,768,0.007407111426194509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,6144,0.02884799904293484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,1024,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,256,0.1650222274992201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,256,0.008240888516108194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,128,0.027999109692043726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,65536,0.47689421971638996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,1536,0.004650666481918759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,32,0.015671110815472074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,4096,0.008411555654472774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,768,0.007842666572994655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,768,0.01941511034965515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,12288,0.05587644709481133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,10240,0.014069333672523499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,512,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,1024,0.004056000047259861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,6144,0.06638933552636041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,1024,0.05262755685382419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,512,0.009908444351620143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,3584,0.054639998409483165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,512,0.009167110754383935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,3584,0.010272888673676385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,5120,0.015684444043371413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,512,0.0359555549091763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,16384,0.04145333170890808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,6144,0.0063591111037466265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,8192,0.04539377821816338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,2048,0.007348444726732042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,16384,1.4450818167792425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,8192,0.016420443852742512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,10240,0.06833955314424303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,51200,0.6875342263115777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,3584,0.03605955508020189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,256,1.6226168738471136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,5120,0.027021333575248718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,5120,0.017418666018380057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,32,0.004287111262480418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,256,0.019973332683245342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,512,0.04042133357789781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,256,0.003160888950030009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,5120,0.6660106447007921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,16384,0.023824888798925612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,8192,0.025156444973415796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,3584,0.09308622280756633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,4096,0.24373777707417807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,2048,0.02071644365787506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,4096,0.10455111000272964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,3072,0.03434400094879998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,1536,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,5120,0.5341422292921278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,51200,0.44483110639784074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,3584,0.06692000230153401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,256,0.0033893332713180115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,3072,0.008287111090289222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,512,0.003690666622585721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,2048,0.008781332936551835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,2560,0.019469334019554984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,64,0.0025395556456512874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,64,0.005740444279379315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,65536,1.8905022939046223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,2560,0.009767110976907942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,768,0.0069288888739215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,8192,0.02017422186003791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,7168,0.43370665444268125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,10240,0.14838133917914495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,128,0.008072000410821702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,3584,0.018280888597170513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,2560,0.027615111735132005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,3072,0.018311111463440787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,12288,3.138170666164822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,12288,0.13497777779897055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,7168,0.055659552415212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,51200,0.06143555376264784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,768,0.013558222187889947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,512,0.0034773333205117118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,2560,0.012135110795497894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,6144,0.12823200225830078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,5120,0.011918221910794577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,65536,0.5581422381930882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,768,0.005625777774386936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,16384,11.669655693901909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,2048,0.1413253280851576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,1024,0.016557332542207506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,2560,0.014933332800865173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,2560,0.013457777599493662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,51200,3.547760857476128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,1536,0.008120888637171851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,2560,0.05204888847139147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,1024,0.003969777789380816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,8192,0.9141306347317166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,16384,0.02534488836924235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,2560,0.019131556153297424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,768,0.01019022199842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,1536,0.083024885919359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,512,0.00925244473748737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,4096,0.006134222365087933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,3584,0.08917333020104302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,5120,0.03424177898301019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,12288,0.01791200041770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,51200,0.07820977767308553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,7168,0.00907466643386417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,32,0.041449778609805636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,5120,0.024544888072543677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,10240,0.08871022197935317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,12288,0.8681199815538195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,2560,0.02894222074084812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,10240,0.03407911128467984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,12288,0.41570221053229434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,128,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,3072,0.01147466649611791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,2560,0.027434666951497395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,12288,0.027701333165168762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,3584,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,64,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,1536,0.031184001101387873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,768,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,512,0.005551110953092575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,2048,0.006343110981914733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,12288,1.7001244227091472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,1024,0.06615377797020806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,3584,0.052652445104387074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,2048,0.013023999830087027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,8192,0.03266133202446832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,3072,0.008372444245550368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,8192,0.01584533353646596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,1536,0.00868266655339135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,8192,0.15709777673085532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,16384,0.01601777805222405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,512,0.007688889073001013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,2048,0.019188443819681805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,128,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,512,0.0073626670572492815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,65536,1.008287959628635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,3072,0.028279112444983587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,256,0.004575110971927643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,8192,0.08496089114083184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,2048,0.07967911163965861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,32,0.008837333156002892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,512,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,32,0.003173333282272021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,768,0.014767110347747803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,6144,0.008254222571849823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,7168,0.007240888973077138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,768,0.009029332962301042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,3072,0.024578667349285547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,5120,0.09815466403961182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,7168,0.023751111494170293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,2048,0.00656355544924736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,1536,0.00499555551343494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,768,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,7168,0.015289778510729471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,768,0.006050666587220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,2048,0.0751439995235867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,1536,0.0105591110057301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,1536,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,51200,0.029870222012201946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,3072,0.04807910985416836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,2048,0.12873155540890163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,8192,0.019663110375404358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,2560,0.007897777689827813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,1536,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,3584,0.010326222413116032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,128,0.019003555178642273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,1024,0.00573600000805325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,4096,0.09585422277450562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,6144,0.020464888877338834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,1024,0.014814222852389017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,128,0.015815110670195687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,768,0.016973333226309884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,3072,0.010436444646782346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,12288,0.0177848885456721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,1024,0.35289777649773496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,4096,0.1546542247136434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,256,0.01167022188504537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,12288,0.0158942225906584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,1024,0.012247111234400006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,3072,0.04254577888382805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,6144,0.027443556321991816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,2560,0.031207111146714952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,256,0.004478222380081813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,2048,0.037167112032572426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,5120,0.009731555150614845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,3584,0.007548444800906711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,12288,1.6822933620876734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,4096,0.054133332437939115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,2048,0.05012444323963589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,256,0.0038924444880750445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,6144,0.008207999997668797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,16384,0.044329779015647046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,12288,0.03464622298876444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,51200,0.05520888831880358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,51200,0.05913866890801323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,256,0.0053280000057485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,32,0.003117333269781537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,2560,0.007722666694058313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,16384,1.0585039986504448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,10240,0.3510631190405951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,16384,2.553715599907769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,10240,0.021969777014520433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,3072,0.10864800214767456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,32,0.0058604445722368025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,12288,0.026540444956885442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,32,0.013841778039932251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,6144,0.015488000379668342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,512,0.013260444833172692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,4096,0.133586671617296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,64,0.005958222266700532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,64,0.032171554035610624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,10240,0.020727111233605277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,8192,0.07245688968234591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,2048,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,7168,0.042740444342295326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,65536,0.4878080156114366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,768,0.06088888645172119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,16384,0.1125519937939114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,3584,0.006744888921578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,3072,0.02225422196918064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,2560,0.022524444593323603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,256,0.0188417782386144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,65536,26.999259101019963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,1024,0.012115555504957834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,128,0.0038497779104444715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,2560,0.05964177846908569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,512,0.01731911136044396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,768,0.048602667119767934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,768,0.006392000036107169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,51200,0.7454684575398763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,512,0.005132444616821078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,768,0.18223110834757486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,768,0.38116176923116046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,256,0.07781510882907443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,4096,0.029778665966457788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,51200,0.5549448861016167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,10240,0.011795555551846823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,51200,1.3855892817179363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,51200,0.04419377777311537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,12288,0.035325331820382014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,1024,0.007325333025720384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,2048,0.009671111073758867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,768,0.055128001504474215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,16384,0.02167111138502757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,64,0.006191111273235745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,65536,0.5497786733839247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,512,0.012823111481136747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,5120,0.028957333829667833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,1024,0.3999351130591498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,16384,0.3553591039445665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,256,0.005107555538415909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,65536,0.35587644577026367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,1536,0.02665511104795668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,128,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,2048,0.027660444378852844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,8192,0.01166488892502255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,2048,0.08675022257698907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,512,0.003685333248641756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,12288,0.043225778473748096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,256,0.004379555583000183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,3072,0.012040889097584618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,16384,0.01364177796575758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,64,0.005229333208666907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,128,0.0042160000238153665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,4096,0.17357067267100015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,3072,2.6701289282904734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,256,0.003629333443111844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,2560,0.022242666946517095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,6144,0.028328001499176025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,1024,0.009867555565304225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,768,0.012418666647540199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,1024,0.009983110758993361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,64,0.002567111204067866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,16384,0.38120444615681964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,768,0.0919155544704861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,2048,0.0055440002017551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,32,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,512,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,64,0.0029022221763928733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,12288,0.2731528811984592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,512,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,8192,0.30405155817667645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,4096,0.00866844422287411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,51200,0.07512088616689046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,8192,0.015459555718633862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,7168,0.018552889426549275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,2560,0.042504890097512134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,64,0.004867555366622077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,3584,0.0945804450247023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,512,0.005703111075692707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,16384,0.023208000593715247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,7168,0.0251511103577084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,2560,0.03614044520590041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,3072,0.012633778154850006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,512,0.0035048888789282907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,16384,0.042825778325398765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,1024,0.004067555483844546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,4096,0.02075199948416816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,256,0.0028897778441508612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,768,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,32,0.003007111036115222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,8192,0.17112710740831164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,1024,0.013468444347381592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,5120,0.04639644424120585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,4096,0.021691555778185528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,3584,0.014999111493428549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,3584,0.0103511114915212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,16384,0.4169813262091742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,12288,0.0720577769809299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,2048,0.005610666755172942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,32,0.016632889707883198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,5120,0.024234667420387268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,6144,0.09067289034525554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,8192,0.0652737749947442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,16384,0.03734577695528666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,2560,0.9917706383599175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,2560,0.04328088959058126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,3072,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,128,0.004372444417741564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,3584,0.3409546746148004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,4096,0.07547466622458564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,2048,0.019800888167487252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,2048,0.015962666935390897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,64,0.0038977778620190094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,32,0.0317537784576416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,6144,0.027845333019892376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,8192,1.6421759923299153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,5120,0.07974755764007568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,32,0.005798222290145026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,2560,0.021474666065639917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,16384,0.042043556769688926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,256,0.016338666280110676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,16384,0.012526222401195101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,64,0.0027840000887711844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,6144,0.1984399954477946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,1024,0.008242666721343994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,3584,0.1430693335003323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,768,0.010875555376211802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,7168,0.02040088839001126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,3584,0.05163910984992981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,6144,0.09145600265926784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,65536,0.2023528946770562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,10240,0.07146400213241577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,3072,0.013470222552617392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,768,0.009924444059530893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,5120,1.7804675632052953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,7168,0.007853333320882585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,32,0.010752889017264048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,3072,0.05297777718967862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,5120,0.012834666503800286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,10240,0.08195466465420194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,1024,0.01908888916174571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,12288,0.2541875574323866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,4096,0.020785777105225455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,256,0.0032088889016045462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,2560,0.012332444389661154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,51200,0.060600890053643115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,3584,1.183370696173774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,256,0.003969777789380816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,3584,0.03946933481428359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,3072,0.013882666826248169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,2048,0.008038221961922115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,256,0.025239111648665533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,12288,0.06688888867696126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,3072,0.025685333543353613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,128,0.005221333354711533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,256,0.013389333254761167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,768,0.07644710938135783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,3072,0.024762666887707178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,3584,0.2808684508005778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,12288,0.015060444672902426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,12288,0.05249777767393324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,2048,0.012318222059143914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,10240,0.6129715707567003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,3584,0.010225777824719747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,3072,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,10240,0.15084977944691977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,12288,0.1310675541559855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,7168,0.05350311266051399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,2560,0.008336889247099558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,65536,0.2379831208123101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,256,0.007923555870850882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,2048,0.1172035535176595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,3584,0.017868444323539734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,16384,0.08956444263458252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,768,0.008477333519193862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,128,0.006636444479227066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,65536,0.6472871038648818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,1536,0.0064088888466358185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,768,0.017988445030318368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,3584,0.22446134355333117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,256,0.002995555599530538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,10240,0.0670880013042026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,768,0.009144888983832465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,8192,0.014912888407707214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,128,0.004364444563786189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,32,0.005104889058404499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,1024,0.200363556543986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,64,0.002699555622206794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,65536,0.12152889039781357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,5120,0.00926755534278022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,8192,0.24480978647867838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,3584,0.09352888663609822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,1024,0.010122666756312052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,3584,0.6101119783189561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,7168,0.006818666640255187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,64,0.009849777652157677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,16384,0.5992417865329319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,7168,0.06614844666586982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,65536,1.410164409213596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,256,1.02165953318278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,128,0.05605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,512,0.20296711391872832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,10240,0.01404800017674764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,2048,0.006904888898134232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,6144,0.0540648897488912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,2560,0.00590133335855272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,5120,2.0565448337131076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,128,0.0069466663731469055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,65536,0.32430844836764866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,51200,0.10524088806576198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,7168,2.1020586225721574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,64,0.008583111067612966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,3072,0.14153599739074707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,65536,19.517871432834202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,4096,0.028909332222408716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,512,0.17230933242373994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,1536,0.007678222325113084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,1024,0.00980355590581894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,5120,0.10759022500779893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,16384,0.28409067789713544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,512,0.024497777223587036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,8192,0.01479288935661316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,3072,0.09604622258080377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,6144,0.04115110966894362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,1024,0.01328000012371275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,128,0.0029173334025674393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,512,0.006903111106819577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,128,0.007802666889296637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,1024,0.018823999497625563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,256,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,8192,0.6511199739244249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,10240,0.119687106874254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,2560,0.07447199689017402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,32,0.005070222334729301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,256,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,7168,0.023294222023752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,7168,0.012127999630239276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,16384,0.3340933322906494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,32,0.0033644443998734155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,2048,0.00620888877246115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,12288,0.05396355523003472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,32,0.06624266836378309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,64,0.07551822397443983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,4096,0.08988444672690497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,64,0.005142222262091107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,512,0.005349333501524395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,64,0.005390222287840313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,2048,0.025897777742809717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,7168,0.014409777190950183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,10240,0.03982222080230713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,4096,0.02613777750068241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,51200,0.8817626635233561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,512,0.039242668284310236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,16384,0.016142222616407607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,1024,0.006468444648716185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,5120,0.014087110757827759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,768,0.008370666868156856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,4096,0.011046222514576383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,256,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,1024,0.015261333849694995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,7168,1.790747536553277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,512,0.011729777687125735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,16384,0.5225928624471029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,4096,0.00849777791235182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,2048,0.008002666963471307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,512,0.015537778536478678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,2560,0.07463466458850436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,7168,0.1033137771818373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,51200,0.36316622628106016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,51200,0.1005564464463128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,1024,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,256,0.004014222158326043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,16384,0.04124622212515937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,16384,0.008672888908121321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,256,0.005102222164471944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,2560,0.008270222279760573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,10240,6.179388258192275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,2560,0.009240888886981541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,3072,0.1507048871782091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,256,0.02709688908523983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,3072,0.02788977821667989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,10240,0.012601777911186218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,6144,0.14181333118014866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,512,0.02913244565327962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,256,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,8192,0.07436622513665093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,5120,0.5105279816521539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,1536,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,512,0.005406222409672207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,32,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,128,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,512,0.005173333403136996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,16384,0.06943288776609632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,3584,0.01127199994193183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,2048,0.009335111412737105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,32,0.8528550995720757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,10240,0.4302506711747911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,512,0.004511111312442356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,2048,0.007616889145639207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,1024,0.09133599864112006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,3584,0.07663733429378934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,8192,0.26877689361572266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,64,0.0069715554515520734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,1536,0.2258195612165663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,3072,0.011405333048767514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,51200,0.0084915558497111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,12288,0.020415110720528495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,512,0.01166133334239324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,16384,0.9063573413425021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,768,0.1633288860321045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,6144,0.012236444486512078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,64,0.02012800011369917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,8192,0.44364089436001247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,128,0.004607111215591431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,16384,0.48949511845906574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,128,0.08353955878151788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,1024,0.008426666259765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,512,0.011077333655622272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,8192,0.030135111676322088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,65536,3.131593704223633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,512,0.01718844473361969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,2048,0.0069395556218094295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,3584,0.191758222050137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,768,0.013996443814701505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,4096,0.03274222215016683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,3584,0.12610133488972983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,8192,0.011573333707120685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,1024,0.010209778116808997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,6144,0.025342222717073228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,256,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,256,0.004459555364317364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,8192,0.24874488512674967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,768,0.004973333328962326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,65536,6.9294238620334205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,10240,0.38584621747334796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,4096,0.011569778124491373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,5120,0.0555057790544298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,65536,7.045552995469834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,6144,0.03792533278465271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,256,0.09377955728107029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,64,0.004109333372778363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,7168,0.04136000076929728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,51200,0.13499377833472356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,2560,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,16384,0.03604622350798713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,16384,0.03985333442687988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,51200,5.641630384657119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,1024,0.026216889421145122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,1536,0.01386044422785441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,12288,0.22287021742926705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,256,0.009586667021115622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,512,0.01662577854262458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,5120,0.012216000093354119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,51200,1.0530995262993705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,16384,0.042599111795425415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,6144,0.018928888771269057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,3584,0.17155199580722383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,2560,0.07083377573225233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,2560,0.008770667016506195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,1536,0.0638151102595859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,1024,0.005124444348944558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,1024,0.009762666291660732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,3584,0.1053804424073961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,65536,0.14146844546000162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,6144,0.013190222283204397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,16384,0.062239110469818115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,64,0.009820444716347588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,6144,0.013770666387346057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,128,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,3072,0.11063466469446818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,64,0.0032808888289663526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,10240,0.10213955243428548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,10240,0.050006220738093056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,2048,0.048192888498306274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,3072,0.03492800063557095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,3072,0.060560888714260526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,12288,0.03865333398183187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,5120,0.06534044610129462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,2560,0.007787555456161499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,8192,3.268972396850586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,5120,0.07926844225989448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,1024,0.00777599960565567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,4096,0.009487999810112847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,6144,1.6722720464070637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,10240,0.019523556033770244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,4096,0.0937804447280036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,2560,0.007910221815109253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,65536,27.283299763997395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,10240,0.02308711078431871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,2560,2.110067579481337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,3072,0.6748391257392036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,512,0.0063999998900625445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,2048,0.034099555677837796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,512,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,64,0.00508622204264005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,6144,0.03680977887577481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,4096,0.02053600052992503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,51200,0.052199999491373696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,4096,0.012778666284349231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,6144,0.012999999854299756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,5120,0.1573235591252645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,51200,6.309032864040798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,3072,0.14847644170125326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,65536,0.6325484381781684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,3072,0.01685244507259793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,10240,0.102256887488895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,512,0.03321866525544061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,65536,0.5323857731289333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,10240,0.152566221025255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,2560,0.009087111386987898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,2560,0.16962844795650908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,256,0.01072177787621816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,1024,0.006216889040337668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,65536,0.4414026737213135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,32,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,64,0.002623999905255106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,3072,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,1024,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,16384,0.037328001525666975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,7168,0.013193777865833707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,16384,0.042994668086369835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,3584,0.030057777961095173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,64,0.007256000406212277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,2048,0.02479022244612376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,8192,0.5037377675374349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,512,0.004528000122971005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,8192,0.06595110893249512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,16384,0.174180441432529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,1024,0.005092444519201915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,32,0.003176888864901331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,65536,0.15309511290656194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,768,0.00620977787507905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,2560,0.016755556066830952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,8192,1.051418622334798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,4096,0.016399110356966656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,12288,0.06073955694834391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,128,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,32,1.3875680499606664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,6144,0.0321324434545305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,3072,0.04247200157907274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,128,0.005513777749405966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,768,0.005803555664088991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,10240,0.01350844403107961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,2560,0.005968889014588461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,1024,0.20336000124613443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,1536,0.011128000087208219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,8192,0.006137777947717243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,512,0.003669333126809862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,10240,0.0318177772892846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,4096,0.020990222692489624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,512,0.0086666668454806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,3584,0.006814222368929122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,12288,0.254859553443061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,5120,0.012688889271683164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,1536,0.057757331265343555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,64,0.0027475555737813315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,10240,0.0239982224173016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,6144,0.2532435523139106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,8192,0.00608800020482805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,51200,0.8920728895399305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,128,0.0033280000918441345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,3584,0.1608017815483941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,65536,0.14265067047542995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,12288,0.006559999866618051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,12288,0.104020443227556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,256,0.15568799442715114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,8192,0.04302666584650675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,128,0.0029608888758553397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,8192,0.006685333533419504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,12288,0.1841822200351291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,10240,0.008084444536103142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,65536,0.8943644629584419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,256,0.00510044437315729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,4096,0.1196204423904419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,7168,0.010738666686746808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,2560,0.008038221961922115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,3072,0.006026666611433029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,3072,0.01904622217019399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,256,0.002769777758253945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,7168,0.013350221845838757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,65536,1.0537502500745985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,2048,0.012448888685968188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,256,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,32,0.004173333446184794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,5120,0.16015377309587267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,128,0.003474666840500302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,32,0.015469332536061605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,3584,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,10240,0.11752711402045356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,7168,0.023818666736284893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,5120,0.013427555561065674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,3584,0.17966311507754854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,768,0.01223111069864697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,1024,0.011735111474990845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,51200,8.687894185384115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,10240,0.3479315439860026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,65536,0.9797431098090278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,512,0.06511111391915216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,6144,0.037681778271993004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,256,0.5175857543945312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,256,0.03720266620318095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,768,0.01706133286158244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,3072,0.10691466596391465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,768,0.01257955531279246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,2048,0.009290667043791877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,51200,0.07298844390445285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,2048,0.030947556098302204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,1024,0.08456977870729233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,512,0.00926399976015091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,1024,0.05060977737108866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,3584,0.04310311211480034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,768,0.044645332627826266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,10240,0.5101564195421007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,768,0.45725780063205296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,128,0.005517333332035277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,8192,0.0343155562877655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,12288,0.04262933466169569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,3584,0.07722222142749362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,12288,5.4404784308539496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,768,0.01242844429281023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,2560,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,512,0.0255386663807763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,512,0.0038604442444112566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,8192,0.026527111728986103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,3072,0.006878222028414409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,768,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,51200,0.07704355319341023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,2560,0.7210648854573568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,16384,0.07623822159237333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,3072,0.008716444174448649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,256,0.005964444329341252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,768,0.012279111478063794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,12288,0.4588809013366699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,128,0.004310222135649787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,10240,0.018539556198649935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,768,0.005559999909665849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,3072,0.009324444664849175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,2560,0.07755200068155925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,5120,2.8268551296657987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,10240,0.026399999856948853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,10240,0.04455022348297966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,64,0.004965333475006951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,3072,0.022291556000709534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,3584,0.31295379002888996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,5120,0.07239911291334364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,10240,0.04190399911668566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,5120,0.010009778042634329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,512,0.6430577701992458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,1536,0.009395555489593083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,64,0.003903999924659729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,3072,0.15769244564904109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,32,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,2560,0.007788444558779399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,256,0.003015999992688497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,2560,0.022231111923853557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,768,0.03519999980926514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,5120,0.011941333611806234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,3072,0.007798222204049428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,3584,0.8968088361952039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,128,0.0038897775941424896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,1024,0.012512000070677863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,768,0.02805866797765096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,128,0.008287111090289222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,6144,0.036228444841172956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,4096,0.015858666764365304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,8192,0.34507022963629824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,16384,0.024784000383483037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,65536,1.605033768547906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,16384,0.19499466154310438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,128,0.002720000015364753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,128,0.00554933316177792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,7168,0.08396622207429673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,51200,2.3287457360161676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,4096,0.03990666733847724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,16384,0.41294132338629824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,4096,0.11346133550008138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,51200,0.5915528933207194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,16384,0.6930649015638563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,16384,0.021594666772418555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,16384,0.13928177621629503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,4096,0.009868444667922126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,12288,0.026527111728986103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,5120,0.029762665430704754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,5120,1.2070817947387695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,4096,0.14703288343217638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,6144,0.07322488890753852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,1536,0.044512000348832875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,65536,0.1275493303934733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,51200,0.09787288639280532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,10240,3.4307448069254556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,32,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,2560,1.9413689507378473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,10240,0.1752871142493354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,3072,0.009896888501114314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,128,0.0035217776894569397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,6144,0.04173688756095039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,4096,0.007052444749408298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,8192,0.016604445046848722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,6144,0.08119022183948092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,512,0.005096000101831225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,1024,0.01719199948840671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,512,0.007080888582600488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,64,0.0036862223512596558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,51200,0.35723643832736546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,768,0.007511110769377814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,2048,0.30216177304585773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,64,0.003261333331465721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,768,0.003640000190999773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,51200,0.7149679925706652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,7168,0.019595555133289762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,1024,0.005872000008821487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,10240,0.2631146642896864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,1536,0.03583466675546434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,1024,0.005534222142563925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,7168,0.12202400631374782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,65536,0.17349421977996826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,10240,0.32868888643052846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,32,0.01953066719902886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,10240,0.04286400145954556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,3584,0.11652444468604194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,256,0.003931555483076307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,51200,0.06451200114356147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,3072,0.011633777783976661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,5120,0.015790222419632804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,3584,0.0177502218219969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,12288,0.13737600379519993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,3584,0.05659733215967814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,8192,0.2660924328698052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,256,0.0412124428484175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,5120,0.06622578038109674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,5120,0.03829244441456265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,256,0.01069866700304879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,8192,0.011876444849703046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,512,0.021661332911915247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,3072,0.009497777455382878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,1024,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,3072,0.020059555768966675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,3584,0.029331554969151814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,1024,0.006287999865081575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,768,0.008519111408127679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,2560,0.007242666350470648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,128,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,6144,0.0059057776298787855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,768,0.016162667009565566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,2048,0.07732799980375502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,5120,0.1437271171145969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,51200,1.2388667000664604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,7168,0.008896888958083259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,768,0.010524444282054901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,4096,1.3034328884548612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,5120,0.018184888694021437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,8192,0.397291554345025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,3584,0.014211555322011312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,128,0.0047084444926844705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,3072,0.032532443602879844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,10240,0.24598489867316353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,2560,0.008663111262851292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,256,0.0047866664826869965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,768,0.00479911102188958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,256,0.004502222355869082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,4096,0.04808888832728068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,1536,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,32,0.006712889091836081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,12288,0.07488977909088135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,256,0.019721777902709115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,5120,4.67555660671658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,32,0.01182933317290412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,4096,0.1184577809439765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,7168,6.197458479139541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,3584,0.010693333215183683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,6144,0.016837333639462788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,16384,0.10671910974714492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,3584,0.1722560061348809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,3584,0.007649777664078607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,32,0.006302222195598815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,32,0.014902222487661572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,65536,0.4614231321546767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,10240,0.009846222069528367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,32,0.008639111287064022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,1024,0.00812977800766627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,10240,0.049999111228519015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,8192,0.03257511059443156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,8192,0.03696177734269036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,512,0.01483022173245748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,5120,0.13138222694396973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,3584,0.01425155500570933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,8192,0.7580666542053223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,512,0.03332533439000448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,8192,0.03873422079616123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,512,0.008553778131802877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,768,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,10240,0.038889778984917536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,512,0.007708444363541073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,7168,0.012210666305489011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,7168,0.48460976282755536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,5120,0.9756284289889865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,1024,0.0977475576930576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,5120,0.007888889147175683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,3584,0.009345778160625035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,64,0.007206222249401941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,1024,0.007409777906205919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,1024,0.006048000107208888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,2560,0.013332444760534497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,6144,0.557602670457628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,1536,0.05390577846103244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,5120,0.024608888559871252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,5120,0.015444444285498725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,256,0.006066666709052191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,128,0.004013333469629288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,64,0.0034560000316964257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,32,0.006070222291681502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,768,0.021612443857722815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,6144,0.12174667252434625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,768,0.009689778089523315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,512,0.01847111185391744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,6144,0.059952888223859996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,7168,0.1835786634021335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,1536,0.027435556054115295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,128,0.004953777624501122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,1536,0.05857155720392863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,256,0.25801155302259654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,3072,0.010859555668301053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,3584,0.3258391221364339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,128,0.0029680000411139596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,6144,0.06820355521308051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,3072,0.05968888600667318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,64,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,1024,0.01295377810796102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,65536,0.37471911642286515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,768,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,1536,0.16677867041693795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,2560,0.008209777375062307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,32,0.0026266665922270883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,512,0.01909955508179135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,7168,0.033210668298933245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,3584,0.013191999660597907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,2048,0.08807288938098484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,2560,0.3668355676862929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,32,0.007038222418891058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,5120,1.2599990632798936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,3584,0.011547555526097616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,65536,0.10560533073213364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,2048,0.009743111001120673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,2048,0.00796888851457172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,1536,0.3845057752397325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,768,0.0041644444896115195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,128,0.0041857775714662345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,10240,0.28405600123935276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,2048,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,10240,0.024686222275098164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,8192,0.13671733273400202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,512,0.0032559999575217566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,1536,0.04497777753406101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,10240,0.08741955624686347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,16384,0.02204977803760105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,6144,0.18579999605814615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,6144,0.019300444258583915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,512,0.005584888988071018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,16384,0.4531048668755426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,256,0.15537777211931017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,8192,1.0144568549262154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,1024,0.03971911138958401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,7168,0.015764445066452026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,51200,0.7343244552612305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,16384,0.02032888929049174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,8192,0.010642666783597736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,128,0.008750222623348236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,16384,0.03770577907562256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,128,0.005186666631036335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,12288,0.020119110743204754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,512,0.0032844444115956626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,16384,0.08139466577106051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,10240,3.0276205274793835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,7168,0.014092443717850579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,2048,0.02588266630967458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,32,0.003391999958289994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,1024,0.011653333074516721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,128,0.027746667464574177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,64,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,8192,0.2241582340664334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,10240,0.039721777041753135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,65536,0.25289866659376353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,65536,0.29986222585042316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,512,0.00903911143541336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,6144,0.01869244376818339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,2560,0.012664889295895895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,768,0.004977777600288391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,12288,0.015145777000321282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,12288,0.06966844532224867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,12288,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,256,0.005321777943107817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,1536,0.005393777870469623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,3584,0.1274684402677748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,2560,0.005752888818581899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,32,0.013533333109484779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,1024,0.030772444274690416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,5120,1.025220447116428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,3072,0.11162577735053168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,4096,0.05722666449016995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,1024,0.006160888820886612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,8192,0.028320001231299505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,7168,0.18575555748409697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,8192,0.020523554748959012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,3072,0.008615111311276754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,1024,0.009549332989586724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,1536,0.024080889092551336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,256,0.04989333285225762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,1024,0.006888000087605582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,768,0.009355555805895064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,16384,0.21279911200205484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,2048,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,5120,0.08367733160654704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,768,0.0047377778424157035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,12288,0.018195556269751657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,5120,0.13044177161322698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,256,0.020853334003024634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,7168,0.013722666435771518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,65536,3.1872054206000437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,512,0.018778666853904724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,16384,0.008341333104504479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,768,0.014382221632533602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,512,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,12288,0.09246044688754612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,10240,0.007796444826655918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,6144,0.011012444065676795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,51200,0.06243377923965454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,768,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,51200,0.11786577436659072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,768,0.02827199962404039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,12288,0.026905778381559584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,2560,0.08872533506817287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,8192,0.6150959862603081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,32,0.03750844465361701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,8192,0.1889288822809855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,768,0.01663733356528812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,3584,0.021035555336210463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,16384,0.017234666479958426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,16384,0.05467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,2048,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,64,0.007385777930418651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,10240,0.022646221849653456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,12288,0.019644444187482197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,512,0.006052444378534953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,128,0.014581332604090372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,10240,3.2165565490722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,8192,0.08620622422960068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,10240,0.3131422201792399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,2048,0.2659502294328478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,768,0.004059555629889171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,8192,0.12597510549757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,5120,0.04926222231653002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,4096,0.00868711123863856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,16384,0.38675467173258465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,256,0.012611555556456247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,12288,0.02660088903374142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,8192,0.022812444302770827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,65536,0.07338755660586886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,3584,0.02390844457679325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,768,0.004948444250557157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,12288,0.3485573397742377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,128,0.004743111216359668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,2048,0.00738933351304796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,7168,0.034845332304636635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,3584,0.03250133328967624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,51200,0.07612978087531196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,12288,0.02597155504756504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,10240,0.02260444396071964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,1024,0.019096000327004325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,1536,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,1024,0.025457777910762366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,10240,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,768,0.22881245613098145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,3072,0.007391110890441471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,16384,0.6827057732476128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,3584,0.9910951190524631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,2048,0.007161777880456712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,2560,0.08516089121500652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,7168,0.02883466747072008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,256,0.02146666745344798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,128,0.0033564445459180405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,1536,0.020103999310069613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,2048,0.09531644317838882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,51200,255.99650065104166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,1536,1.059543079800076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,6144,0.04149422380659316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,4096,0.05756444401211209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,4096,0.05340444379382663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,768,0.07094311051898532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,10240,0.014940443966123792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,8192,0.026329777306980554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,3072,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,65536,0.08053600125842623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,768,0.0877457790904575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,64,0.010227555202113258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,128,0.01316088851955202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,256,0.01740800009833442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,1024,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,1536,0.03735111157099406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,2560,0.04056266612476773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,6144,0.031326221095191106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,64,0.0027884443600972495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,12288,0.028132445282406274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,256,0.0052764444715446895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,8192,0.011915555430783166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,768,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,6144,0.27109066645304364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,3072,0.01701866587003072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,256,0.005903111149867375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,768,0.013791999883121915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,8192,0.48201778199937606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,10240,0.13923733764224583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,51200,0.04273777869012621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,256,0.4118506643507216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,1536,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,3072,0.009865778187910715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,5120,0.08050933149125841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,2560,0.012716444830099741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,65536,0.10435555378595988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,2560,0.29410134421454537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,51200,0.11051199833552043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,4096,0.019109333554903667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,768,0.006312888943486744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,1536,0.006063999815119638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,64,0.03758755657407973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,2048,0.010620444185203977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,32,0.053189333942201406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,10240,0.1935573286480374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,65536,0.21783999601999918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,4096,0.1381831169128418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,2048,0.03669688767857022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,3584,0.30258666144477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,512,0.007553777760929531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,12288,0.21532266669803193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,768,0.02979555394914415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,32,0.0038399998512532976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,512,0.009149333669079674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,768,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,3584,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,5120,1.8505795796712239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,3584,0.031055112679799397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,1536,0.009883555273214975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,8192,0.024466666910383437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,8192,0.014416888356208801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,65536,0.8583200242784289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,1024,0.005630222045713001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,128,0.006193777753247156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,1536,0.06981155607435438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,8192,0.020632889535692003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,1536,0.008776889079146916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,1024,0.007072889142566257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,1024,0.02219999995496538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,12288,0.02735288937886556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,7168,0.028132445282406274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,8192,0.39624977111816406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,512,0.0273297776778539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,1024,0.01421955558988783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,4096,0.15672088993920222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,51200,0.1284773349761963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,3072,0.030828442838456895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,32,0.008422222402360704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,16384,0.1368968884150187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,3584,0.01796800063716041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,8192,0.018774222996499803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,51200,2.3913724687364364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,6144,0.12138933605617946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,8192,0.03982933362325033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,2048,0.030058665408028498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,7168,0.059102223979102246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,3072,0.05983377827538384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,3584,0.050733334488338895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,2560,0.005583999885453119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,12288,0.17284533712599012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,51200,2.4901919894748263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,5120,0.05273600088225471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,6144,0.2101893292533027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,65536,1.0385546154446073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,5120,0.03920977645450168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,16384,0.028806222809685603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,16384,0.23643912209404838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,65536,1.9501849280463324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,4096,0.01595288846227858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,16384,0.028759999407662287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,3584,0.039817776944902204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,64,0.01464355488618215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,3584,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,64,0.008559111091825698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,8192,0.01202044470442666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,12288,0.4310106701321072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,5120,0.0103502223889033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,16384,0.23611911137898764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,51200,0.10384355651007758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,12288,0.13429689407348633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,2560,0.007536888950400882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,512,0.003495111233658261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,512,0.025077333052953083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,2560,0.010302222437328762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,1536,0.060243553585476346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,128,0.00507377791735861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,5120,0.05994133154551188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,2560,0.06470755736033122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,2048,0.012568888564904531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,65536,0.7792133225335015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,5120,0.03736622134844462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,6144,0.040602667464150324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,10240,0.023448000351587932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,12288,3.695574018690321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,64,0.009777777724795872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,51200,0.06861155562930636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,8192,0.23310934172736275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,4096,0.05188622077306112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,2560,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,1536,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,51200,0.05730488565233019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,1024,0.010283555421564314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,512,0.011128889189826118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,32,0.019552888141738046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,3584,0.1761031150817871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,2048,0.007265778051482306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,4096,0.022993778189023335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,64,0.013034666577974955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,10240,0.04609244399600559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,7168,0.018384888768196106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,51200,0.26794934272766113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,768,0.003608889049953885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,3584,0.030064000023735896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,512,0.07676621940400866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,4096,0.1035991112391154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,10240,0.04033599959479438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,3072,0.05936355723275078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,32,0.006077333456940121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,3584,0.018223999275101554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,768,0.13642577330271402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,8192,0.22481332884894478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,2560,0.008214222060309516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,1024,0.009832888841629028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,51200,0.024784000383483037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,16384,0.04048177931043837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,8192,0.025777777036031086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,16384,0.5904809104071723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,51200,1.2761573791503906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,51200,0.014041778114106921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,65536,0.21029955810970732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,768,0.00630044440428416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,3584,0.008609777523411645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,51200,4.990951114230685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,32,0.020765332712067496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,256,0.15721244282192653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,2048,0.0070186663005087115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,256,0.005670222143332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,1024,0.009974222216341231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,1536,0.006480000085300869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,4096,0.14340089427100286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,1024,0.08872710996203953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,8192,0.032814221249686346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,2560,0.021294222937689886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,2560,0.007359111474619971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,64,0.008088000118732452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,128,0.016626665989557903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,2048,0.05592177973853218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,64,0.002577777744995223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,6144,0.010297777752081553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,1536,0.010065777434243096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,51200,1.0998871061537001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,6144,0.03421599997414483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,3584,0.011397333608733283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,51200,1.4232035742865667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,3584,0.01351377781894472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,6144,0.04645599921544393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,2048,0.012802666260136498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,4096,0.023553777072164748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,10240,0.030592888593673706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,2560,0.18369600507948133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,10240,0.02400711178779602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,3072,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,51200,0.26030755043029785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,10240,0.013587555951542325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,256,0.010466666685210334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,12288,0.025938666529125635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,3072,0.07594222492641874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,65536,0.36304177178276914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,512,0.028910222980711196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,2048,0.008369777765538957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,6144,0.024999111890792847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,8192,0.02195555633968777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,3584,1.2223866780598958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,3072,0.010500444306267632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,64,0.031962666246626116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,8192,0.007768889268239339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,65536,3.8665084838867188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,5120,0.0794959995481703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,10240,0.04505955510669284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,16384,0.32205155160692006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,256,0.005323555734422471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,1024,0.003924444317817688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,10240,0.6260462337070042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,8192,0.03293777836693658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,3072,0.019638222124841478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,65536,0.15183199776543513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,2048,0.01404355631934272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,8192,0.3017635610368517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,512,0.00562755556570159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,3072,0.015795555379655626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,4096,0.021299555897712708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,64,0.06281866629918416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,64,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,3584,0.017084444562594097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,768,0.004005333201752769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,8192,0.011787555283970304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,3072,0.012850667039553324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,51200,0.04760444495413038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,128,0.03128711051411099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,10240,0.06744444370269775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,51200,2.903047137790256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,4096,0.07397688759697808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,64,0.005094222310516569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,12288,0.0147733340660731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,64,0.02868266569243537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,8192,0.05279466509819031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,64,0.0998248921500312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,1024,0.01145066652033064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,8192,0.031888888941870794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,10240,0.016024889217482675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,6144,0.011624000138706632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,5120,0.013008000122176277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,2048,0.01702933344576094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,5120,0.010962666736708747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,768,0.005768888940413793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,2560,0.013000888956917657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,2048,0.00941422250535753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,51200,5.965855068630642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,3072,0.01094044413831499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,4096,0.012333333492279053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,512,0.007893333004580604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,16384,0.031494223409228854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,1024,0.009722666607962714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,10240,7.330507914225261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,2560,0.010420444111029306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,16384,0.03496977686882019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,512,0.042141334878073804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,32,0.0035546666218174826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,16384,0.0839608907699585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,1536,0.006575111299753189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,2560,0.00758844448460473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,16384,0.7997173203362359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,8192,0.026569777064853247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,1024,0.03433510992262099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,10240,0.03479377759827508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,3584,0.9988524119059244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,16384,4.992371453179254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,512,0.009027555584907532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,1536,0.041518221298853554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,128,0.008359111017651027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,256,0.04141777753829956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,256,0.12623022662268743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,128,0.004567111117972268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,64,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,12288,0.10540355576409234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,3584,0.08928355243470933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,2560,1.248520851135254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,5120,0.023504000571038988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,4096,0.01055555542310079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,6144,0.06218666500515408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,3584,0.05397599935531616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,64,0.004111999852789773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,12288,0.08586488829718696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,4096,0.017645332548353408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,256,0.021938666701316833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,32,0.009748444788985783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,64,0.0038480001191298165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,6144,0.07454399930106269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,4096,0.027777777777777776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,3072,0.02938488788074917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,256,0.06436533398098417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,10240,0.019664888580640156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,65536,2.1665403578016496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,768,0.2551857895321316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,3072,0.012151111331250934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,12288,0.045231110519833036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,1024,0.019392000304328073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,10240,0.02181688944498698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,1024,0.008534222013420528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,8192,0.06254044506284925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,64,0.002604444407754474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,10240,0.35704178280300564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,1536,0.004840888910823398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,7168,0.07808978027767606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,64,0.002630222174856398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,16384,0.07525866561465792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,8192,0.20521956019931367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,12288,0.04937511020236545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,2560,0.015001777145597668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,2560,0.005635555419656966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,10240,0.045054223802354604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,4096,0.013946667313575745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,65536,0.42832088470458984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,51200,0.39399109946356875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,512,0.0041555555330382455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,512,0.0052915554907586836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,1536,0.027635556128289964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,3072,0.07813688781526354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,256,0.005174222091833751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,1024,0.004228444563017951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,7168,0.13726754983266196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,3072,0.04768622252676222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,1024,0.0354728897412618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,10240,0.02647200061215295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,32,0.003177777760558658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,1024,0.009071110851234859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,1536,0.005663110978073544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,32,0.11662488513522679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,6144,0.017326222525702584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,10240,0.0362408889664544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,128,0.004292444636424382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,65536,0.2096488873163859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,8192,2.1004666222466364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,16384,0.012329777909649743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,6144,0.058522668149736196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,51200,13.028861999511719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,65536,0.12434932920667861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,1024,0.0047128887640105355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,7168,0.5698506567213271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,512,0.012364444633324942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,2560,0.008722667064931657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,1024,0.014114666316244336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,1536,0.01798311041461097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,51200,0.46638933817545575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,10240,0.029701332251230877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,12288,0.01218488895230823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,2560,0.04083555605676439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,16384,0.24186844295925566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,12288,0.035283555587132774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,4096,0.01570577753914727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,768,0.003499555504984326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,512,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,6144,0.007041778001520369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,1024,0.003999111139112049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,7168,0.02665688925319248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,8192,0.06418844726350573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,128,0.005521777603361342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,256,0.007594666547245449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,64,0.02014488809638553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,65536,1.5062115987141926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,10240,0.017443555924627516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,65536,0.1763786607318454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,5120,0.11354045073191325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,768,0.019145778483814664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,8192,0.8027493158976237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,6144,0.03279733326699998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,4096,0.0140337778462304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,2560,0.00868800034125646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,1024,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,64,0.009577777650621202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,768,0.00609777785009808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,7168,0.9848835203382703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,5120,0.01165600038237042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,3072,0.009376000199053023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,3072,0.04755466514163547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,2048,0.009230222139093611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,2560,0.010968888799349466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,1024,0.008685333033402761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,7168,0.6333893140157064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,65536,3.936830308702257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,128,0.004976888911591636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,512,0.010121777653694153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,256,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,64,0.015092444088723926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,6144,0.019151111443837483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,128,0.004039111236731212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,3072,0.012065778175989786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,2560,0.008045333127180735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,6144,0.523639996846517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,1536,0.0053004444473319584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,2560,0.024596444434589807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,64,0.004736888739797804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,256,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,128,0.004044444610675176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,10240,0.015876443849669564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,512,0.004716444346639845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,2048,0.033745778931511775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,4096,0.07020622491836548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,768,0.008757332960764566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,51200,2.0299947526719833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,768,0.017842666970358956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,12288,0.17693155341678193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,1024,0.013735111388895245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,2048,0.005554666535721884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,8192,1.2400684356689453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,16384,0.18101333247290718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,1536,0.11445511711968316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,4096,0.03945599993069967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,768,0.02859555681546529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,51200,0.08033155732684664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,2048,0.0970568921830919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,7168,0.6183048884073893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,5120,0.11843821737501357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,65536,1.0343128840128581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,51200,0.6634808646308051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,8192,0.0314319994714525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,4096,0.07112266620000203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,16384,0.09969511297014023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,6144,0.14621867073906794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,6144,0.02947288751602173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,2048,0.015980444020695157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,128,0.002752888947725296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,128,0.009217778013812171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,128,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,7168,0.024697777297761705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,6144,0.021513778302404616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,10240,0.02970400121476915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,51200,0.49946313434176975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,12288,0.4023626645406087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,64,0.0028168888141711554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,4096,0.006311111152172089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,64,0.002628444383541743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,5120,0.006794666664467917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,8192,0.13858221636878118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,128,0.010111110905806223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,8192,0.006542222367392645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,768,0.021876444419225056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,51200,0.2705013222164578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,51200,0.20581155353122285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,32,0.004621333132187526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,4096,0.06144266658359104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,51200,0.06696622239218818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,1024,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,256,0.0053013331360287136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,6144,0.036750220590167575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,8192,0.0464915566974216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,8192,0.035700443718168474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,1536,0.008030222521887885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,4096,0.011885333392355176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,12288,0.050755557086732656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,8192,1.285095108879937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,7168,0.013824889229403602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,3584,0.06691822078492907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,1024,0.030457778109444514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,5120,0.06773777802785237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,6144,0.11160977681477864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,1024,0.00566133318675889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,64,0.0026400000270869997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,51200,0.07058222426308526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,2048,0.00685777763525645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,3072,0.2828835646311442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,10240,0.02701866626739502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,256,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,768,0.0037671112351947357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,1024,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,128,0.002951111023624738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,1536,0.027968890137142603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,4096,1.3413759867350261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,2560,0.009696889254781935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,3072,0.02130311065249973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,512,0.01386044422785441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,1536,0.03911822372012668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,10240,0.01351377781894472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,51200,0.009691555466916826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,6144,0.006380444599522485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,1536,0.007721777591440413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,2560,0.010021333065297868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,2560,0.010687111152542962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,16384,0.11912977695465088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,3072,0.027311111489931744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,51200,0.011703111231327057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,12288,0.034154666794670954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,1024,0.08391377660963271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,5120,0.008647999829716152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,64,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,5120,0.017800889081425138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,5120,0.01072355525361167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,256,0.020440888073709276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,7168,0.009425777528021071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,4096,0.012880889077981314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,768,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,8192,0.8278711107042102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,4096,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,16384,3.1320525275336375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,7168,0.09027644660737778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,10240,0.022501332892311945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,1024,0.005943111247486538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,5120,0.052520000272327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,1024,0.16553866863250732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,16384,0.20339643955230713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,1024,0.30392445458306205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,16384,0.033329778247409396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,1536,0.01128000020980835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,128,0.009551111194822524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,12288,0.02051199972629547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,65536,1.4588133494059246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,3584,0.014764444695578681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,768,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,1024,0.05238222082455953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,7168,0.13794044653574625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,3072,0.04586844311820137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,1536,0.00849955528974533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,10240,0.032842665910720825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,3584,0.017628444565667045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,10240,0.24415821499294707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,768,0.010777778095669217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,1024,0.023183110687467787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,256,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,7168,0.022312000393867493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,8192,0.034407999780442976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,3584,0.008094222181373173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,7168,0.2545359929402669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,1536,0.020784000555674236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,128,0.0029528888149393927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,64,0.010051555931568146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,2560,0.11009599765141805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,256,0.010671999719407825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,32,0.005719111197524601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,3072,0.18105955918629965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,3584,0.014555555250909595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,2048,0.006816888848940532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,3584,0.28059199121263295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,1536,0.006514666808976068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,2048,0.03536799881193373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,65536,7.120554606119792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,12288,0.03976710968547397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,128,0.0031351111829280853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,6144,0.02658133374320136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,256,0.005980444451173146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,5120,0.018406222263971966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,2560,0.03253688746028476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,3584,0.016552888684802584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,2560,7.219874911838108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,768,0.0443795555167728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,32,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,512,0.00777511133088006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,65536,1.9181323581271703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,256,0.018020444446139865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,65536,45.183325873480904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,128,0.03864622116088867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,1536,0.06870400243335299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,2560,0.055460443099339805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,7168,2.5866639879014754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,256,0.004061333421203825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,12288,2.586225721571181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,512,0.06218577755822075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,10240,0.1566231118308173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,12288,0.0465262234210968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,128,0.011734222372372946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,64,0.007821333077218797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,2048,0.0679360032081604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,256,0.04089066717359755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,5120,0.014101333088344999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,7168,0.01850488947497474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,12288,0.22536177105373809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,32,0.007189333438873291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,10240,0.4332622157202826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,12288,0.061631997426350914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,4096,0.01125955581665039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,10240,0.03658488723966811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,7168,0.01926399932967292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,2560,0.014279110564125909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,12288,0.07862844732072619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,10240,0.2965288956960042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,6144,0.010911111202504901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,1024,0.13319110870361328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,3584,0.033378667301601835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,7168,0.5564364327324761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,16384,10.685748630099825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,256,0.008570666942331526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,3072,0.019533332851197984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,16384,0.47869157791137695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,5120,0.017789334058761597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,16384,0.0306951105594635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,12288,0.34150134192572695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,16384,0.14346488316853842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,3584,0.02440888848569658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,1536,0.01720977822939555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,16384,0.02834577692879571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,2048,0.3611626625061035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,51200,2.573752933078342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,16384,0.05615999963548449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,512,0.028956443071365356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,256,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,7168,0.5227946705288357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,12288,0.12411022186279297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,32,0.19780977567036948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,12288,0.01851377718978458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,1024,0.012578667038016848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,2560,0.04533866710133023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,16384,0.44851112365722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,1536,0.04768800073199802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,768,0.007268444531493717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,2048,0.007370666497283512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,256,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,768,0.02884799904293484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,2048,0.019744889603720773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,5120,0.012140444583363004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,10240,0.0273617770936754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,3072,0.878053347269694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,51200,1.6402995851304796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,128,0.005355555564165115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,7168,0.5254302024841309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,65536,4.817412482367621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,16384,0.10740088754230076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,4096,0.1591013272603353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,3584,0.07923110988405016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,1024,0.04656444324387444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,1024,0.004664888812435998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,3072,0.1899200015597873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,128,0.0027955555253558685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,16384,0.28149867057800293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,51200,0.048271109660466514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,256,0.005646222167544895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,6144,0.0096133334769143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,2560,0.005702222386995952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,1024,0.009584888815879822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,10240,1.913257810804579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,256,0.03704799877272712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,6144,0.19674489233228895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,10240,0.05696355634265476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,7168,0.13200978438059488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,3072,0.03738222188419766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,2048,0.007229333122571309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,64,0.0027022223091787765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,10240,0.10276266601350571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,2048,0.011316444310877057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,256,0.012748444245921241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,768,0.052552888790766396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,3072,0.010666666759385003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,768,0.5108595424228245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,2048,0.39217156834072536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,10240,0.0395084453953637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,16384,0.025048888391918604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,1536,0.009368889033794403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,4096,0.008193777667151557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,1024,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,5120,0.028565333949195013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,16384,0.9958178202311198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,2560,0.012149333126015134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,1024,0.015473778049151102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,1024,0.12678399350908068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,2048,0.007280888656775157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,3584,0.02254755629433526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,12288,0.012706666356987424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,51200,1.198828485276964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,2560,0.7075262069702148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,64,0.008037333687146505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,7168,0.1887777778837416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,6144,0.16753599378797743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,10240,0.2793644534216987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,12288,0.04110844598876105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,12288,0.23758310741848418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,8192,0.012484444512261285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,4096,0.013861333330472311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,1024,0.054126222928365074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,4096,0.027970665031009253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,1024,0.0058266665372583605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,7168,0.017669333351982962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,1536,0.33376267221238876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,10240,0.014578666951921252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,512,0.007125332951545715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,512,0.007121777368916406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,12288,0.021764443980322942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,2560,0.014711111783981323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,51200,0.08462755547629462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,5120,0.032327112224366926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,5120,0.12282578150431316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,51200,3.1727903154161243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,3584,0.014946666028764514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,512,0.0064737776087390045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,1536,0.00906844437122345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,1536,0.008692444198661381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,6144,0.026778666509522334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,4096,0.27074400583902997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,3584,0.017131555411550734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,256,0.0047333331571684945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,3584,0.008826666408114964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,64,0.03985777828428481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,65536,0.30676089392768013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,16384,0.09043111403783162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,12288,0.021498666869269475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,2048,0.006773333168692059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,2048,0.01184266640080346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,768,0.013526221944226159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,16384,0.023920888702074688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,65536,0.18712711334228516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,64,0.006203555398517185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,768,0.21435466077592638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,64,0.004005333201752769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,3072,0.01351466692156262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,2048,0.006337777607970768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,2048,0.009933333430025313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,51200,0.21543021996816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,1024,0.021054221524132624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,2560,0.007629333270920648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,1536,1.9557804531521266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,10240,0.2693191104465061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,64,0.003264888914095031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,128,0.00927466650803884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,51200,0.06672355863783094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,32,0.0028986665937635633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,8192,0.1994160016377767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,3072,0.007856888903511895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,3072,0.010085333552625444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,8192,0.21346488263871935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,64,0.00272088891102208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,3072,0.013153778182135688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,16384,0.34006577067905003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,6144,0.028201778729756672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,768,0.1507520013385349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,32,0.007541333635648091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,2048,0.6605271233452691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,12288,0.024272888898849487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,16384,0.4499697685241699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,1024,0.007139555282062954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,1536,0.005976888868543837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,128,0.02037244372897678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,7168,0.052729778819613986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,8192,0.08780799971686469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,8192,0.05580266979005602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,16384,0.10469777716530694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,5120,0.22771822081671822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,3584,0.013576000101036496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,10240,1.5287502076890733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,32,0.004583999928500918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,128,0.01834844383928511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,6144,0.027601778507232666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,6144,0.0181057784292433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,2560,0.1980124447080824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,512,0.009309333231714036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,3584,0.011247999966144562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,128,0.004679999831649992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,3072,0.008275555239783393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,10240,19.571631537543404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,2560,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,512,0.01938577824168735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,1024,0.003964444415436851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,5120,0.11718666553497314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,6144,0.01872711049185859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,3584,0.007745777567227681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,64,0.005176888985766305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,2560,0.006520888871616787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,32,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,8192,0.050662222835752696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,7168,0.06460266643100314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,128,0.005720888988839255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,8192,0.02494844463136461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,12288,0.025154666768179998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,12288,0.02493688960870107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,128,0.003077333379122946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,8192,0.013834666874673633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,12288,0.01576622161600325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,2560,0.013231111069520315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,256,0.005161777966552311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,65536,6.761041853162978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,12288,0.036213331752353244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,4096,0.01885333326127794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,10240,0.028804444604449805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,2048,0.04920533299446106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,3072,0.03050133254792955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,256,0.013406222065289816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,4096,0.5848568810356988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,512,0.006722666737106111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,51200,0.11033778058158027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,256,0.005299555758635203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,8192,0.039495110511779785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,12288,0.09761244720882839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,2560,0.017083555459976196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,2560,0.009101333717505137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,6144,0.009733333355850643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,6144,0.036370668146345354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,128,0.0028159999185138275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,51200,0.16242310735914442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,3584,0.17692000336117217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,6144,0.02049777739577823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,512,0.007823111282454597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,2560,0.07084177600012885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,6144,0.023349333140585158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,1024,0.005600000007285013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,8192,0.061363557974497475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,2048,0.017608889275126986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,1536,0.0057706667317284485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,2560,0.21159022384219697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,2048,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,1536,0.06336444616317749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,32,0.005902222047249476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,51200,0.06226133637958103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,8192,0.02632355524433984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,3072,0.007746666669845581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,12288,0.104184885819753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,6144,0.22133066919114855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,128,0.004472889006137848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,7168,0.02013066742155287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,7168,0.013912000589900546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,65536,0.37121955553690594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,4096,0.019056000643306308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,256,0.006073777874310811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,2048,0.009356444080670675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,512,0.003133333391613431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,8192,0.3777004347907172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,51200,0.430181344350179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,7168,0.02560088866286808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,6144,0.009959999885823991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,64,0.028335111008750066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,128,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,65536,0.09724799791971843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,10240,0.03670488794644674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,65536,2.088650597466363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,1024,0.027053332991070215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,8192,0.36618932088216144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,12288,0.02165688905451033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,32,0.013355555633703867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,51200,0.04251289036538866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,1024,0.005250666704442766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,256,0.037717332442601524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,7168,0.12708532810211182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,65536,0.999235577053494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,51200,0.07487110959159003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,6144,0.01370933320787218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,1536,0.007567110988828871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,2048,0.00723555518521203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,3072,0.10930399762259589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,5120,0.023762666516833838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,16384,0.3085164493984646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,4096,0.05831288629108005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,32,0.002941333378354708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,3072,0.045724444919162326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,7168,0.1217128833134969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,2560,0.012680889003806643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,256,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,4096,0.02632177703910404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,512,0.014828445182906257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,2048,0.015658666690190632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,32,0.00258933338854048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,4096,0.07422933313581678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,10240,0.016148444679048326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,16384,0.23142666286892363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,51200,14.896752251519096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,768,0.00925333301226298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,256,0.005473777651786804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,12288,0.07496888769997491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,2048,0.10088000032636855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,32,0.0031973332580592898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,128,0.0028613333900769553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,1024,0.009174221919642555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,64,0.0028124445428450904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,51200,1.4764098061455622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,51200,0.06429777542750041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,3072,0.009556444154845344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,128,0.006217777729034424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,4096,0.06001777781380547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,6144,0.014755555325084262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,8192,0.03120000163714091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,512,0.0351200004418691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,8192,2.771066665649414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,128,0.006613333192136552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,256,0.006552000012662675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,10240,0.027357333236270483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,6144,0.012418666647540199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,1536,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,5120,0.030838221311569214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,5120,0.03973866502443949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,10240,0.014665777484575907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,5120,0.09318133195241292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,512,0.0032720000793536506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,4096,0.12313777870602077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,64,0.005172444300519096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,256,0.005352889084153705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,3584,0.02253599961598714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,51200,0.06948177682028876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,512,0.05880710813734266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,1024,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,65536,26.400680541992188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,5120,0.06784800026151869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,2048,0.0507315562831031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,10240,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,256,0.0030862221287356485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,16384,0.024674667252434626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,4096,0.016022221909628976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,6144,0.04427466789881388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,2048,0.11258488231235081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,64,0.005769777629110549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,16384,0.13523288567860922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,51200,0.09930044412612915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,8192,0.010519111322032081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,6144,0.016529776983790927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,512,0.025853332546022203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,768,0.003777777983082665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,12288,6.108918931749131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,1536,0.07505866554048327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,2560,0.013977777626779346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,3584,0.009330666727489894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,1024,0.01293688929743237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,12288,0.23525688383314344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,512,0.006996444529957241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,32,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,1024,0.0057884446448749965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,7168,4.829837375217013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,256,0.01982133256064521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,4096,0.3695351017846002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,1536,0.03236088818973965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,51200,0.3798853291405572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,1024,0.02115999990039402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,7168,0.010802666346232096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,2048,0.09490844276216294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,2560,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,51200,0.16740443971421984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,768,0.020167110694779288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,5120,0.008371555142932469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,5120,0.07244444555706449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,4096,0.3319937653011746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,256,0.003083555648724238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,2048,0.007592889169851939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,128,0.017665778597195942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,3584,0.00797155582242542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,10240,0.03302399979697333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,1024,0.005803555664088991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,5120,0.0310533344745636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,32,0.007189333438873291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,128,0.005285333428117964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,3584,0.5182639757792155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,2048,0.019380443625979953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,3072,0.04087466663784451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,10240,0.024248000648286607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,768,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,2048,0.296615997950236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,3584,0.2420142226748996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,32,0.019674667053752475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,65536,0.17176000277201334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,3072,0.012088889049159156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,64,0.024752888414594863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,3584,0.9063279893663195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,6144,0.02369333306948344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,4096,0.011190222369299995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,2560,0.012221333053376941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,32,0.0030862221287356485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,65536,0.08415733443366157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,16384,0.022874666584862605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,768,0.45713334613376194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,12288,0.8180924521552192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,65536,0.7161377800835503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,2560,0.009416888985368941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,2560,0.013306666579511432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,51200,0.03719288773006863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,1024,0.010919111470381418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,1024,0.006140444427728653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,2048,0.00777511133088006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,16384,0.05612888601091173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,12288,0.030272887812720403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,4096,0.01789866718981001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,1024,0.06007644202974108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,2048,0.15712800290849474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,128,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,768,0.017080888152122498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,12288,0.009221333596441481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,6144,0.030606223477257624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,65536,0.21870756149291992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,768,0.005895111295912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,6144,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,7168,0.02459466622935401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,256,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,2048,0.03590133455064561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,10240,0.3578337828318278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,5120,0.011671110987663269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,768,0.05233866638607449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,768,0.03429066803720262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,16384,0.5431084632873535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,512,0.009038222332795462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,128,0.02086044516828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,5120,0.09475377533170913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,2560,0.013242666920026144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,256,0.009002666506502364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,8192,0.18387111028035483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,2048,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,5120,0.04177155428462558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,4096,0.3366764386494954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,768,0.046104888121287026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,8192,0.0646631121635437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,6144,0.20496711466047499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,512,0.014387556248241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,1536,0.05019199848175049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,1024,0.023210666245884363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,3584,0.015951999359660678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,16384,0.09049777852164374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,12288,0.021016889148288306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,7168,0.020844444632530212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,7168,0.02199733257293701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,4096,0.9654142591688367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,1024,0.007045333584149678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,768,0.03103289008140564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,1024,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,6144,0.00720711135201984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,6144,0.055065777566697865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,8192,0.020262221495310467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,2048,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,3584,0.006725333217117522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,2048,0.1612844467163086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,16384,0.0878373318248325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,256,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,3072,0.01591999994383918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,6144,0.009906666974226633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,3584,0.04932266804907057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,64,0.002562666725781229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,2048,0.037294222248925105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,1024,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,64,0.01613066593805949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,7168,0.016977777083714802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,5120,0.058892442120446094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,1024,0.004036444342798657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,51200,0.11383999718560113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,256,0.011901333100265928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,7168,0.2432462109459771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,1024,0.0041751112374994485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,3584,0.05216799841986763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,1024,0.00508088908261723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,51200,0.06555289030075073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,6144,0.03328266739845276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,10240,0.017108443710539076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,12288,0.03011200163099501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,5120,0.011226666470368704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,65536,0.360841777589586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,6144,0.12381777498457168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,6144,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,10240,0.0743102232615153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,2560,0.05717955695258247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,1536,0.008620444271299574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,3072,0.00889866633547677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,1024,0.003963555726740095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,16384,0.1005022194650438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,1024,0.005800888770156437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,2048,0.02318488889270359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,512,0.011394666300879585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,4096,0.27106666564941406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,5120,0.014876445134480795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,64,0.0026444445053736367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,12288,9.063690185546875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,2560,0.547586652967665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,5120,0.023131555981106226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,256,0.022077333596017625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,2560,0.04801155461205376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,8192,0.38055910004509824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,1536,0.00814222213294771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,3072,0.016382222374280293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,1536,0.007725333174069722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,16384,0.6682204670376248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,128,0.008937777744399177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,768,0.039797332551744245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,256,0.03210577699873183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,32,0.01202933324707879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,32,0.02274755636850993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,12288,1.0292204750908744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,6144,0.012274666792816587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,128,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,5120,0.20563911067114937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,7168,0.25126756562127006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,3584,0.049398223559061684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,1024,0.06646933158238728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,3072,0.08063111040327284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,128,0.004313777718279096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,1024,0.009930666950013902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,1536,0.009297777381208207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,256,0.007636444436179266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,2560,0.09547199805577596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,1024,0.005306666509972678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,1536,0.0396364430586497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,5120,0.07285155852635701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,8192,0.016385777129067313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,10240,0.1655662192238702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,256,0.005376889059940974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,8192,0.02737155556678772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,12288,0.3844790988498264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,51200,2.9808737436930337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,51200,0.3879208829667833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,512,0.014223110344674854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,2560,0.049711998966005116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,65536,5.2209014892578125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,7168,0.013023999830087027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,128,0.019863999552196927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,8192,0.021534222695562575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,2048,0.022873777482244704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,3072,1.0084595150417752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,768,0.03433955709139506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,4096,0.012337778177526263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,768,0.005156444592608346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,65536,1.6955155266655817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,51200,0.34469244215223527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,64,0.003064888839920362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,4096,0.0740764472219679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,3584,0.05659821960661146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,8192,0.2881680064731174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,3584,0.011135111252466837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,1536,0.04541422261132134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,7168,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,65536,0.1848968929714627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,10240,0.36254578166537815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,5120,0.5836702452765571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,6144,0.007506666911972894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,8192,0.02197422252760993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,65536,0.07929333051045735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,1024,0.04223733478122287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,768,0.03228444523281521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,2048,0.01329955541425281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,10240,0.12681510713365343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,7168,0.009111999637550777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,64,0.002858666703104973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,10240,0.023277334041065637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,5120,0.010607110957304636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,10240,0.06758755445480347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,10240,0.020398222737842135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,256,0.006414222220579783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,2048,0.015563555889659457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,256,0.00526311124364535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,7168,0.02020711037847731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,128,0.004960888789759742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,2560,0.022328888376553852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,65536,0.7962844636705187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,5120,0.025545777546034917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,16384,0.054846223857667714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,4096,0.03885866536034478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,512,0.01311288856797748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,4096,0.013609777722093793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,512,0.44049599435594344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,3072,0.0932240022553338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,512,0.0034942221310403612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,51200,0.613734245300293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,2560,0.00981155534585317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,7168,0.14964000384012857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,5120,0.0829555524720086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,768,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,8192,0.3701306713951959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,2560,0.011247110863526663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,1024,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,7168,0.01611288885275523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,128,0.027066666218969557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,12288,0.18581155935923258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,256,0.004752000172932942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,256,0.010420444111029306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,4096,0.03324000040690104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,51200,0.03578222129080031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,1536,0.06059200233883328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,3072,0.11041777663760716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,128,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,256,0.008912000391218398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,3584,0.2433084381951226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,256,0.005087999833954705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,128,0.01089777797460556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,51200,0.050309333536359996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,1536,0.011162666810883416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,256,0.01761155492729611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,2560,0.007002666592597961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,5120,0.016230222251680162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,1024,0.0637484457757738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,2560,0.01556622154182858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,16384,0.08953955438401963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,65536,0.08576533529493545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,3584,0.007784888976150089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,51200,0.20072621769375273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,12288,0.011583999627166323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,16384,0.21766577826605904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,32,0.012682666381200155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,10240,0.01740977830357022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,7168,0.11603555414411758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,64,0.0026595555245876312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,12288,0.0813937783241272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,64,0.002554666664865282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,768,0.02420622275935279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,128,0.008287111090289222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,768,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,3584,0.04569955666859945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,5120,0.21440088748931885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,65536,0.16805599795447457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,7168,0.09088355302810669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,32,0.002491555487116178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,12288,0.015023999743991427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,4096,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,3584,0.13752710819244385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,6144,0.19335556030273438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,2560,0.0057555557125144536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,3072,0.07152089145448473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,65536,2.761621263292101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,4096,0.02515022291077508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,10240,0.6046266555786133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,51200,0.06229777468575371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,512,0.0033111110743549135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,4096,0.8818737665812174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,7168,0.09336000018649632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,3072,0.051128000020980835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,65536,6.096581353081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,10240,0.02437244521247016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,2048,0.022111111217074927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,6144,0.02992977698644002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,12288,0.07456444369422065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,5120,1.3877333535088434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,3584,1.249375131395128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,7168,0.014144889182514615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,4096,0.007469333708286285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,64,0.0034799998005231223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,1536,0.013258666627936892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,4096,0.2550942103068034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,128,0.00847466703918245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,2560,0.11449066797892253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,1024,0.01183733344078064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,6144,0.04124266571468777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,3072,0.08724621931711833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,2560,0.015955555770132277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,256,0.0294773346847958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,4096,0.5373839802212185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,3072,0.008091555701361762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,65536,0.1862106720606486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,256,0.004317333300908406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,4096,0.1156417793697781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,64,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,128,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,2560,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,51200,0.8447528945075141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,1536,0.07724533478418986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,256,0.005021333280536863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,5120,0.3164844512939453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,10240,0.02348711093266805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,12288,0.017138666576809354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,3072,0.05582311418321398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,128,0.010302222437328762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,1024,0.00526311124364535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,5120,0.1765902174843682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,768,0.023652444283167522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,5120,0.009180444810125563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,768,0.0038204445607132385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,2560,0.008421333299742805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,12288,0.37435823016696507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,512,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,5120,0.14080088668399388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,768,0.005872888697518243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,51200,0.16561422083112928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,128,0.004555555681387584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,16384,0.23504355218675402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,128,0.00637511122557852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,8192,0.05347111158900791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,5120,0.031381332212024264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,5120,0.16493777434031168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,2560,0.17946222093370226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,65536,0.08052533202701144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,3072,1.8396125369601781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,2560,0.009119999905427298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,2048,0.5788879924350315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,2560,0.021368000242445204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,16384,0.5241680145263672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,7168,0.019282667173279654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,2560,0.07773600021998088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,4096,0.1695715586344401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,7168,1.968597412109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,16384,0.0926240020328098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,2560,0.10271377695931329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,128,0.004173333446184794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,256,0.0033119999700122406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,5120,0.07854933208889432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,12288,0.01645155582163069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,7168,0.035926222801208496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,16384,0.03842133283615112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,2048,0.05645510885450575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,8192,0.021528888079855178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,16384,3.946380615234375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,32,0.04608711269166735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,256,0.006979555719428592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,2560,0.021632888250880774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,768,0.13968000147077772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,10240,0.018914666440751817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,2560,0.017813333206706576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,6144,0.011495999991893768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,51200,3.2094169192843967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,1536,0.017782222893502977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,3072,0.1795039971669515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,2048,0.28050043847825795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,3584,0.014984000060293408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,512,0.029623111089070637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,768,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,768,0.00924800005224016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,5120,0.033732444047927856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,3072,0.009018667042255402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,256,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,5120,0.022222222553359136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,6144,0.009550222092204625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,1536,0.014303111367755465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,2560,0.025827555192841425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,64,0.05269955595334371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,512,0.06872000296910603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,2048,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,1536,0.013908444179428948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,768,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,128,0.004980444494220945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,51200,2.546478271484375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,10240,0.10644355747434829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,1536,0.008226667013433244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,768,0.4611546728346083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,4096,0.08316533433066474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,768,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,5120,0.01718222267097897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,32,0.011340444286664328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,6144,0.02010577751530541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,3072,0.013756444056828817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,1024,0.03197066651450263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,128,0.0321724447939131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,3072,0.023494222097926672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,6144,1.2894951502482097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,6144,0.2330248885684543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,12288,0.09969511297014023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,128,0.016048888365427654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,2048,0.15396800306108263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,8192,0.12453333536783855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,6144,0.011408888631396823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,2560,0.007851555943489075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,2560,0.06676266590754192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,51200,0.05085688829421997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,12288,0.09803999794854058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,10240,6.001462300618489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,7168,0.0447617769241333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,4096,0.019323555959595572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,256,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,256,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,2560,0.013823111024167804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,3584,0.05866310993830363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,2048,0.027803556786643133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,3072,0.04543200135231018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,256,0.004295111116435793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,65536,0.19837155607011583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,4096,0.015351111690203348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,32,0.0064995553758409284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,768,0.006027555714050929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,3584,0.015263110399246216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,768,0.009946666657924652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,2560,0.007676444119877285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,3584,0.008811555802822113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,12288,0.032812443044450544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,2048,0.006735111276308696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,256,0.007348444726732042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,1536,0.12477866808573405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,2048,0.20353955692715117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,8192,1.7616782718234594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,10240,0.22961778110928008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,51200,21.176539103190105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,1024,0.45689598719278973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,3584,0.0750560031996833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,256,0.013573333621025085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,2048,0.017596445149845548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,256,0.010783111055692038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,8192,0.09133066733678182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,2560,0.05602222018771701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,512,0.013439999686347114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,3584,0.32542310820685494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,3072,0.007811555431948767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,3072,0.016674665941132438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,128,0.009292444421185387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,4096,0.008256889051861232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,256,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,512,0.008657777474986183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,128,0.02495199938615163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,64,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,10240,0.027246221899986267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,3584,0.021202666891945735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,512,0.00674044465025266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,3584,0.014528888795110913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,16384,0.027866668171352808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,2560,0.008078222473462423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,12288,0.1265653371810913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,6144,0.019125334090656705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,768,0.007522666619883643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,3072,0.013169777890046438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,768,0.007693332930405934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,51200,0.1538355615403917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,5120,0.05992266866895887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,16384,0.03901066713862949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,32,0.005327111317051782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,3072,0.01365066650840971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,3584,0.006849777781301075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,1024,0.010469333165221743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,16384,0.06226311127344767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,64,0.005514666852023866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,128,0.00739555557568868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,768,0.03973689013057285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,2048,0.016623111234770883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,128,0.005092444519201915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,4096,1.7126746707492406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,12288,0.0967893335554335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,1536,0.06044266621271769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,5120,0.20850043826633027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,12288,0.18271022372775605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,2560,0.09056622452206081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,5120,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,3584,0.00797155582242542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,8192,0.023440000083711412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,256,0.007516444557242923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,512,0.005239999956554837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,512,0.02980088856485155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,12288,0.01834844383928511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,256,0.012533333566453723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,51200,0.16858311494191489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,10240,0.016760889026853774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,1536,0.009639110830095079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,128,0.0029333333174387612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,51200,1.0132097668117948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,10240,0.08290933238135444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,8192,0.013992889059914483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,1024,0.010371555884679159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,65536,1.1020470725165474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,12288,0.032623999648623996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,4096,0.022870222727457683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,51200,0.0850035548210144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,51200,0.4802417755126953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,7168,0.1385893291897244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,512,0.0037671112351947357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,5120,0.02530577778816223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,128,0.0058559998869895935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,256,0.02595111065440708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,768,0.009090666969617208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,4096,0.01608000033431583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,32,0.0047306666771570844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,16384,0.023973332511054143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,64,0.1816106637318929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,256,0.023418666587935552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,1024,0.009282666775915358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,10240,0.051148444414138794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,6144,0.021359999974568684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,32,0.00942133367061615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,4096,0.05788711044523451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,4096,0.008356444537639618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,5120,0.14976978302001953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,128,0.0027280000762807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,8192,0.02832711074087355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,8192,0.02236266599761115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,3584,0.014415999253590902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,512,0.3819013436635335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,128,0.0029786665820413162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,10240,0.06703111198213366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,6144,1.7949884202745225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,1024,0.09691022502051459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,12288,0.0377111103799608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,256,0.01689511040846507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,16384,0.8644933170742459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,512,0.05126133229997423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,1536,0.0195795562532213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,16384,0.024860444996092055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,6144,0.010219555762079027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,1024,0.0038417776425679526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,10240,0.006276444428496891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,5120,0.04439999990993076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,6144,0.19946489069196913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,8192,1.5909484227498372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,512,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,8192,0.019056000643306308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,3584,0.40982667605082196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,8192,1.0238222546047635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,65536,0.5559457672966851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,65536,0.7387724452548556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,2048,0.09420622057384914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,2048,0.03157155712445577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,64,0.005047111047638787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,1024,0.02106133268939124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,51200,0.338626676135593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,768,0.023769777682092454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,768,0.005391110976537068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,3072,0.03353066576851739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,12288,0.01723022262255351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,768,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,1536,0.010334221853150262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,1536,0.00591466658645206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,5120,0.11662667327457005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,256,0.013413333230548434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,2560,0.017514665921529133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,5120,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,7168,0.017165333032608032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,65536,0.6867902013990613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,768,0.006062222023804982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,3584,0.007831110722488826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,10240,3.2602462768554688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,8192,0.052912887599733144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,6144,0.008374222450786168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,3072,0.0075839997993575195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,12288,0.006045333213276333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,1024,0.014171555638313293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,128,0.00546488869521353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,51200,1.140800052218967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,8192,0.014840889308187695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,4096,0.02756266627046797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,64,0.005922666854328579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,1536,0.0064328888224230866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,51200,0.448928885989719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,64,0.017274666163656447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,2560,0.01873333255449931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,2048,0.007231999602582719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,768,0.005107555538415909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,5120,1.6315893597073023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,8192,0.11143378416697185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,6144,0.009836444424258338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,12288,0.025960889127519395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,3072,0.018382221460342407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,65536,0.00814400033818351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,1536,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,128,0.006623111251327727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,512,0.005682666682534748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,4096,0.03363644414477878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,12288,0.015286222100257874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,3584,0.02551288902759552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,12288,46.296610514322914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,1024,0.009910222556855943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,2560,0.008166222108734978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,3584,0.02193244463867611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,8192,0.04935555656750997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,3584,0.09857600265079075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,2560,0.031719999180899724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,10240,0.30003200636969674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,65536,0.18230222331153023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,32,0.005889777921968036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,5120,0.01239822225438224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,12288,0.020960888928837244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,10240,0.02878755662176344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,128,0.012469333079126147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,8192,0.10299466715918647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,10240,0.03879377908176846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,65536,2.0057955847846136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,2560,0.018837332725524902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,7168,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,1024,0.007350222104125553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,10240,0.054535110791524254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,1536,0.04370222157902188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,4096,0.028240889310836792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,12288,0.026576000783178542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,64,0.01627111103799608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,3584,0.13340177800920275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,8192,0.6956622335645887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,7168,0.4519946840074327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,128,0.0027137777457634607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,32,0.003478222423129611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,512,0.005719111197524601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,1536,0.008485332959228093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,8192,0.46227020687527126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,12288,0.9598621792263455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,8192,0.09626133574379815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,51200,0.8473599751790365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,12288,0.027237332529491846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,64,0.004330666528807746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,8192,0.013306666579511432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,51200,0.4211991098192003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,10240,0.11781599786546494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,65536,0.2978915638393826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,7168,0.07123200098673503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,16384,0.4362213346693251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,65536,0.5620026588439941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,128,0.010272888673676385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,8192,0.33152445157368976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,4096,0.07529155413309734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,16384,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,16384,0.021565333008766174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,8192,0.08554310931099786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,512,0.004517333375083076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,5120,0.5721573299831814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,10240,0.0567528870370653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,1536,0.02310933338271247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,32,0.003124444435040156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,10240,0.01555999947918786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,3584,0.01903377804491255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,3584,0.060177776548597545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,3072,0.04747200012207031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,12288,0.5788017908732096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,512,0.13379732767740884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,512,0.04572888877656725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,128,0.14692621760898167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,256,0.017511111166742113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,512,0.022221333450741235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,1536,0.6895982424418131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,256,0.004279110994603899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,6144,0.05506044295099047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,1536,0.056925336519877114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,768,0.004338666796684265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,1024,0.11354222562577988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,1536,0.008998221821255153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,8192,0.04562044474813673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,3584,0.01739911072784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,2048,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,512,0.021134222547213238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,512,0.05611199802822537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,10240,0.01420088940196567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,128,0.005967999911970562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,3584,0.0896400014559428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,1536,0.015772443678643968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,6144,1.1182577345106337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,65536,1.241130616929796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,10240,0.03286400106218126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,512,0.012950222525331708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,65536,0.15895733568403456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,768,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,6144,0.19713777965969512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,512,0.008412444757090675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,8192,0.23112445407443574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,16384,0.2558071083492703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,5120,0.03815911213556925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,12288,0.014126222994592456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,16384,0.03151466780238681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,256,0.012780444489585029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,65536,0.4589013523525662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,128,0.0026453334010309647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,256,0.0037439999481042228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,512,0.0034346667428811393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,7168,2.813205295138889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,51200,1.0269226498074002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,7168,0.0441599984963735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,1024,0.050999999046325684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,3584,0.01276177747382058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,512,0.004588444613748127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,768,0.02957688768704732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,6144,0.025263999899228413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,7168,0.035712000396516584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,512,0.016352888610627916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,7168,0.02012888921631707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,32,0.002647111192345619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,16384,0.03901422354910109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,8192,0.2772515614827474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,6144,0.036496887604395546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,6144,0.012970666918489667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,1024,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,12288,0.20903378062778047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,32,0.008910222185982598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,128,0.0031111112071408164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,1024,0.004524444540341695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,768,0.004143110993835661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,4096,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,32,0.011189333266682096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,4096,0.011397333608733283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,8192,1.050170686509874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,51200,1.2546808454725478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,4096,0.026094221406512793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,64,0.0027502222607533136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,12288,0.08030133114920722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,2560,0.24716088506910536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,3072,0.017089777522616915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,51200,0.0652631123860677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,32,0.005055111315515307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,6144,0.022879999544885423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,10240,0.2553937700059679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,3584,0.00831288927131229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,1536,0.08239377869500054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,16384,0.29143110911051434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,768,0.2801315519544813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,3072,0.014708444476127625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,8192,0.015426667200194465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,5120,0.012640000217490725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,7168,0.27441689703199595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,3072,0.0229759994480345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,3072,0.05979644589953952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,512,0.0035013332962989807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,128,0.028669334120220605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,512,0.01564800077014499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,3072,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,128,0.0027964444210131965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,1536,0.011206222077210745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,32,0.022080888350804646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,2560,0.028752889898088243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,1024,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,256,0.05852355559666952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,512,0.005163555757866966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,7168,0.02961777647336324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,4096,0.02367111047108968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,6144,0.042585776911841504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,1024,0.062330669826931424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,1024,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,768,0.004302222281694412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,51200,0.05328266819318136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,256,0.003363555504216088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,12288,0.01923377811908722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,12288,0.2941608958774143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,3072,0.06200977828767565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,65536,0.5375982390509711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,5120,0.02587822245226966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,16384,0.03391289048724704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,16384,0.6743946605258517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,3072,0.008013333711359236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,5120,0.02443733314673106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,768,0.008055999875068665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,2560,0.008785777621799046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,3072,0.16404267152150473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,2048,0.01586577792962392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,5120,0.09293599923451741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,128,0.0520631108019087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,3072,0.8988168504503039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,1024,0.0682835578918457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,6144,0.03586488962173462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,3072,0.008967999782827165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,4096,0.0877351098590427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,16384,0.02557955516709222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,2048,0.026317333181699116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,2560,0.06758044163386027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,12288,0.021158221695158217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,256,0.008440000315507254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,3072,0.007691555553012424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,10240,0.028004444307751123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,6144,0.329398234685262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,12288,0.147450672255622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,2560,0.010233778092596266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,2048,0.007419555551475949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,8192,0.12478844324747722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,128,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,10240,0.015763555963834126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,5120,0.013905778527259827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,12288,0.04100622071160211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,2560,0.013220444321632385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,2560,0.02977422210905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,256,0.003104888937539525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,32,0.008208889100286696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,32,0.01770666738351186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,10240,0.9067288504706489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,2560,0.013268444273206921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,16384,4.81257078382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,3584,1.3020150926378038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,64,0.014485332700941296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,8192,0.03669688767857022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,16384,0.00637155564294921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,3584,0.5271120071411133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,5120,0.017162667380438913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,51200,0.4797840118408203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,65536,0.6341564390394423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,10240,0.00962488849957784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,256,0.0040835556056764396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,51200,0.008782222039169734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,1536,0.005113777601056629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,3072,0.008446222378147973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,5120,0.030386666456858318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,51200,1.3804737726847331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,8192,0.024907555845048692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,5120,0.1278124385409885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,6144,0.01569688816865285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,12288,0.16093956099616155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,12288,0.029849777619043987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,5120,0.07928711175918579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,10240,0.013609777722093793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,10240,0.08759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,12288,0.021149333980348375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,32,0.0029013332807355454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,3072,0.056187556849585645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,12288,0.008957333034939235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,32,0.006628444625271692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,3072,0.008818666968080733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,16384,0.10238755411571926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,4096,0.11152444945441352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,5120,0.012624888784355588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,8192,0.014140443669425117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,768,0.1252408954832289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,12288,0.03365777929623922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,65536,4.6110454135470915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,16384,0.02500266664557987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,51200,1.3847867117987738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,3072,0.06221066580878364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,3072,0.23541598849826387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,128,0.004585777719815572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,5120,0.11284889115227593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,3584,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,5120,0.02347999976740943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,512,0.004489777816666497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,2048,0.011176889141400656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,256,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,1024,0.027651555008358423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,1536,0.006543111056089401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,5120,0.03382844395107693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,5120,0.0211173329088423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,3072,0.030360000001059637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,3072,0.01754755609565311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,3072,0.009876444107956356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,3072,0.04176000091764662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,3072,0.06506844361623128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,8192,0.1063164472579956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,51200,0.15310133828057185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,512,0.04081155525313483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,5120,0.19949866665734184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,6144,0.03536622060669793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,6144,0.0790826678276062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,128,0.018959111637539335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,3072,0.005928888916969299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,12288,0.022781333989567224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,6144,0.017071111334694754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,8192,1.1264524459838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,8192,0.4090249008602566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,3072,0.009497777455382878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,16384,0.05947911077075534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,12288,0.10531644688712226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,1536,0.052335109975602895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,4096,0.2383955584632026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,512,0.00350222239891688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,64,0.007289778027269576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,16384,0.09965688652462429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,32,0.007139555282062954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,768,0.004943999979231092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,2048,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,8192,0.08072711361779107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,51200,0.04354666670163473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,7168,0.051764445172415845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,51200,2.908128950330946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,2048,0.05597688754399618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,7168,0.03609422180387709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,6144,0.022218666142887537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,51200,2.8233680725097656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,128,0.003702222059170405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,10240,0.028684443897671167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,128,0.005607999861240387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,128,0.03716444306903415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,6144,0.01297511077589459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,512,0.031187554200490315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,2560,0.014888889259762235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,51200,0.2576071156395806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,128,0.11122132672203912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,512,0.006879999819729064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,5120,0.31884177525838214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,3584,0.13379822836981878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,8192,0.4407511022355821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,3072,0.06163022253248426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,256,0.004648888690604104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,7168,0.04243288768662346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,256,0.006063111126422882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,3072,0.6558346748352051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,16384,0.04306044512324863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,1536,0.019488000207477145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,768,0.0036524443162812125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,12288,0.015186667442321777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,3584,0.03433422247568766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,51200,0.36686844295925564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,65536,0.017253332667880587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,512,0.0034257777863078644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,65536,0.7883804639180502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,51200,19.858506944444446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,2560,0.06433333290947808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,2560,0.009242666264375051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,8192,0.01593244406912062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,65536,0.14138933022816977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,3072,0.014161777165200977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,2048,0.04452622267935011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,51200,0.0118231111102634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,16384,0.07955111397637261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,65536,0.3547066582573785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,1024,0.008951110972298516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,10240,17.509402804904514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,128,0.009229333036475712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,1024,0.05956266985999214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,32,0.007376000285148621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,8192,0.06377244657940336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,12288,0.5026879840426975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,8192,0.056448002656300865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,128,0.09496177567376031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,1536,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,128,0.005927111125654644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,1536,0.011017777853541903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,12288,0.03184977836079068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,64,0.0027173333283927706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,1024,0.01089511149459415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,4096,1.1779253217909071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,64,0.00462755560874939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,8192,0.3322382238176134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,7168,0.12467200226253933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,51200,0.28911111089918345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,65536,0.11699199676513672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,2048,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,1024,0.004785777793990241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,1536,0.028185778194003638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,128,0.004097777936193678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,128,0.0052062223354975385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,12288,0.008358221915033128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,512,0.0041200001206662925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,4096,0.013167999684810638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,10240,0.03106844425201416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,8192,0.2891457875569661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,768,0.009604444106419882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,1536,0.009339555270142024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,64,0.002591111179855135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,2048,0.04521955715285408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,51200,4.1052674187554254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,3072,0.03380444314744737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,2560,0.01623644431432088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,64,0.0031431110368834603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,1024,0.00960355583164427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,8192,0.014671110444598727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,6144,0.022848000129063923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,8192,0.03141866789923774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,5120,0.03779200050565932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,1024,0.15333510769738093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,1536,0.009474666582213508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,7168,0.025075554847717285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,128,0.007346666521496243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,7168,0.44442221853468156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,5120,0.029924445682101782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,2560,0.008190222084522247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,3584,0.04277866747644213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,1536,0.006245333287451003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,51200,11.850793626573351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,12288,0.030055999755859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,7168,0.05102044343948364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,768,0.008805333740181392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,4096,0.015233778291278414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,5120,0.013444444371594323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,512,0.014967999524540372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,12288,0.49843645095825195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,3584,0.013694222602579327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,8192,0.3589902189042833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,8192,0.060587558481428355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,7168,0.011759111450778114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,5120,0.009628444082207149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,5120,0.006167111297448476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,10240,0.49530755148993594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,128,0.003168888803985384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,3072,0.03777599996990628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,10240,0.06719555457433064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,8192,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,128,0.018748443987634447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,2048,0.22725155618455675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,65536,0.01831999917825063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,6144,0.18332089318169487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,4096,0.020644444558355544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,3072,0.12383643786112468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,10240,0.11979822317759196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,65536,0.08039110898971558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,4096,0.020917332834667627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,32,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,12288,0.13431644439697266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,512,0.013713777893119387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,512,0.011686222420798408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,6144,0.024313777685165405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,2048,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,3584,0.00907911111911138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,1024,0.003949333396222857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,768,0.003638222399685118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,51200,0.050460444556342236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,256,0.017991110682487488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,2560,0.013216888739003075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,8192,0.006520000182920032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,3584,0.014896889527638754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,2560,0.04288977715704176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,5120,0.01364177796575758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,32,0.0026435556097163092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,51200,5.114991929796007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,5120,0.029322667254341975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,6144,0.030301332473754883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,51200,0.9016062418619791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,32,0.021631111701329548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,3584,0.05705510907702976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,51200,0.07148800293604533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,4096,0.0925368865331014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,512,0.004308444344335132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,2560,0.19269600179460314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,10240,0.01997511088848114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,10240,0.025233777032958135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,5120,0.011986666255527072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,512,0.014516444669829475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,768,0.005952000204059813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,768,0.022652443912294176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,51200,0.04970577690336439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,32,0.005148444324731827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,768,0.010282666318946415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,51200,0.3716053432888455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,16384,3.1364924112955728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,256,0.2074986696243286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,7168,0.14206933975219727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,12288,0.17238222228156197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,10240,0.016389333539538913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,3072,0.10358666711383396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,8192,0.1926791138119168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,1536,0.011797332929240333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,128,0.004480888860093223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,16384,0.7912524541219076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,16384,0.03139288889037238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,256,0.004200889004601372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,16384,0.050831112596723765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,768,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,65536,3.1928399403889975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,4096,0.09026755226982965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,16384,0.0844951139556037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,32,0.007367110914654202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,65536,16.45062425401476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,128,0.01975644462638431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,10240,0.3998888863457574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,512,2.825743145412869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,10240,0.37909600469801163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,128,0.005730666634109285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,512,0.004406222038798862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,3584,0.011533333195580376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,10240,0.03386488887998793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,65536,2.2135963439941406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,256,0.019156444403860304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,51200,0.07743822203742133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,768,0.006091555373536215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,12288,0.03610488772392273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,5120,0.02347911066479153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,768,0.0995111134317186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,768,0.005484444399674733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,2048,0.17224355538686117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,51200,0.1584951082865397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,2560,0.054061333338419594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,2048,0.006075555665625467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,2048,0.01830666595035129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,3584,0.027265777190526325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,6144,0.016522667474216886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,2560,0.00924533357222875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,512,0.004271111140648524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,3584,0.014907555447684394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,5120,1.6369928783840602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,256,0.012853333519564735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,512,0.0034328889515664843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,3584,0.008346666892369589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,12288,0.14871999952528211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,256,0.004104888687531154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,16384,0.32461688253614635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,51200,14.618808322482638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,3072,0.04926933182610405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,3584,0.40521955490112305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,2048,0.017464000317785475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,6144,1.31951109568278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,1536,0.008246222303973304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,512,0.027245332797368366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,51200,0.049376888407601245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,4096,0.043756445248921715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,5120,0.12072622776031494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,3072,0.02880000074704488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,1024,0.012620444099108378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,128,0.005046222358942032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,4096,0.01812088820669386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,65536,0.6871751149495443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,7168,0.36549690034654403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,768,0.004087111188305749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,51200,0.23722667164272734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,65536,2.6345083448621964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,2048,0.08074577649434407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,12288,0.03512710995144314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,256,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,64,0.002560888934466574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,10240,0.3887013329399957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,768,0.003588444242874781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,2048,0.007281777759393056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,1024,0.02956533432006836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,10240,0.1832186645931668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,65536,0.37163911925421816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,8192,0.506299548678928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,2048,0.09485689136717056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,2560,0.060659554269578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,8192,0.668759134080675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,768,0.008416000339719985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,10240,0.09095822440253364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,2048,0.041858666472964816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,64,0.016332444217469957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,51200,0.0682506693734063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,12288,0.2509537802802192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,1536,0.006514666808976068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,65536,0.016349333855840895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,7168,0.3608080016242133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,3584,0.12345333894093831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,5120,0.3060364458296034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,10240,0.0937840011384752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,7168,0.020177778270509508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,5120,0.0762702226638794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,5120,0.02612888813018799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,1536,0.00813600007030699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,128,0.016720889343155753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,1536,0.34200355741712785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,512,0.003488000068399641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,2048,0.013328889177905189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,2560,0.008151999778217739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,65536,2.1309839884440103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,2048,0.02844088938501146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,32,0.007309333317809635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,1024,0.08831466568840875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,12288,0.08422755532794529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,2048,0.006970666348934174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,4096,0.008429333567619324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,65536,0.08514933453665839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,1536,0.01109688894616233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,3072,0.024485333098305598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,2560,0.008196444147162968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,4096,0.08854933579762776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,2048,0.11214666896396214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,768,0.009297777381208207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,65536,1.4000790913899739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,3584,0.1293706629011366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,7168,0.02125155594613817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,7168,0.011053333679835001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,5120,0.02751022246148851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,12288,0.14930489328172472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,2560,0.011346666349305047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,5120,0.18407377931806776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,2560,0.01942311061753167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,4096,0.1833662192026774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,2560,0.03208177619510227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,6144,0.03161955542034573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,512,0.00490133340160052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,4096,1.2872471279568143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,2560,0.02510844502184126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,256,0.01075644459989336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,16384,0.2989191214243571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,10240,0.0773235559463501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,7168,0.3024204572041829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,3072,0.07842666572994657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,8192,2.4418578677707248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,65536,0.4710391362508138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,128,0.19615377320183647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,5120,0.033858666817347206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,5120,0.37221421135796445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,2560,0.015649777319696214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,7168,0.014918223023414612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,3072,0.008879111044936711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,256,0.017858665850427415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,512,0.01590577761332194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,12288,0.18200443850623238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,64,0.016744888491100736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,1024,0.009181333084901174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,768,0.016131555040677387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,10240,0.017332444588343304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,10240,0.12734844949510363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,10240,0.38658756679958767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,16384,0.025959110922283594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,12288,0.027787556250890095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,5120,0.01722044414944119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,10240,0.024193776978386775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,7168,0.012339555554919772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,12288,0.1055235531595018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,1024,0.005000888887378904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,7168,0.014912888407707214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,256,0.007605333295133378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,2048,0.07865244150161743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,64,0.0029306666304667792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,512,0.01292622254954444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,10240,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,16384,0.025379555092917547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,1024,0.005782222168313132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,512,0.005630222045713001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,128,0.005037333402368757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,512,0.004320000194840961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,256,0.006861333217885759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,3584,0.012958221965365939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,65536,0.06890755229526095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,128,0.0026506665680143568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,51200,0.1362640062967936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,2048,0.0073902226156658595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,128,0.044792887237336904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,65536,15.828104654947916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,4096,0.012350222302807702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,12288,0.0891466670566135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,10240,0.041350222296184964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,2048,0.04682933290799459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,512,0.018025777406162687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,10240,0.10826400253507827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,256,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,512,0.003530666646030214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,12288,2.245139651828342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,3072,0.021531555387708876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,6144,0.15387466218736437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,51200,0.06078133318159315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,512,0.005576000031497743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,2048,0.022300443715519373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,1024,0.01680355601840549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,10240,0.10498933659659491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,6144,0.011745777395036487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,2560,0.019164444671736825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,16384,0.00906577789121204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,2560,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,512,0.006114666660626729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,10240,0.01367377738157908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,768,0.010239111052619087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,512,0.006940444310506185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,5120,0.06836622291141085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,5120,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,512,0.0770346654786004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,3584,0.009642666412724389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,256,0.015847111741701763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,12288,0.02387911081314087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,512,0.06255910793940227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,7168,0.06493955850601196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,128,0.006654222392373615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,16384,0.01670488880740272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,2048,0.007831110722488826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,7168,0.013075555364290873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,256,0.005239999956554837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,512,0.00599644457300504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,768,0.009500444763236577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,4096,0.31271288130018443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,4096,0.006265777680608962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,128,0.005371555685997009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,1024,0.005124444348944558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,6144,0.00620977787507905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,3584,0.023336889015303716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,64,0.00600000015563435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,8192,0.025215999947653875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,8192,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,8192,0.10659022463692559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,512,0.009167999857001835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,51200,5.487123701307509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,4096,0.054655998945236206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,7168,0.07406844695409139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,512,0.004809777769777509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,51200,13.327673170301649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,2560,0.04486044579082065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,768,0.00997688869635264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,16384,0.044194665220048696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,12288,9.158033582899305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,512,0.004383111165629493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,768,0.014662222729788886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,2560,0.012774222426944308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,4096,0.017514665921529133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,768,0.018138666947682697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,64,0.029480000336964924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,768,0.019229332605997723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,6144,0.14410311645931667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,768,0.0063999998900625445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,8192,0.06405333015653822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,2560,0.06176533301671346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,6144,0.32487466600206166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,4096,0.009935999910036722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,2048,0.06259466542137994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,768,0.010071111222108206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,2048,0.015598222613334656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,8192,0.04141599933306376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,4096,0.015378667248619927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,3584,0.025913778278562758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,64,0.004997333304749595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,1536,0.006206222292449739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,10240,0.021377777059872944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,2560,0.007144889069928064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,5120,0.01983644399378035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,768,0.3043777677747938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,6144,0.009508444203270806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,16384,0.006720888945791457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,7168,0.06474133332570393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,2560,0.032024890184402466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,5120,0.029935999049080744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,16384,0.022290666898091633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,1024,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,16384,0.571705765194363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,10240,0.08662666877110799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,256,0.02241511146227519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,1536,0.006055111272467508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,32,0.0058746664888328975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,2560,0.018664888209766813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,4096,0.02825066778394911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,3072,0.014475555883513557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,768,0.008524444368150499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,1536,0.008298666940795051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,2048,0.005084444251325396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,768,0.004553777890072929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,3584,0.0324577788511912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,1536,0.07081244389216106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,51200,4.786975860595703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,4096,0.794356452094184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,7168,0.3415519926283095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,1536,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,1536,0.022732444935374792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,32,0.038498666551378034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,6144,0.053470220830705434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,3072,0.009249777429633671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,2048,0.009543110926946005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,2048,0.0084915558497111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,8192,0.03385244475470649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,512,0.026874666412671406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,51200,0.008436444732877944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,2048,0.0058648888435628675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,2560,0.3737395604451497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,128,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,6144,0.04145866632461548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,256,0.11096266905466716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,256,0.03192089001337687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,8192,0.01682222220632765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,1024,0.006076444354322221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,768,0.0034231110993358824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,6144,2.5647786458333335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,12288,0.5682177543640137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,7168,0.03660533494419522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,7168,0.2127759986453586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,6144,1.373654259575738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,7168,0.22001422776116264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,1536,0.011471110913488599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,3584,0.030642665094799463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,2560,0.008815111385451423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,32,0.0031902222997612427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,1024,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,8192,0.02595022155178918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,3072,0.00812977800766627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,2560,0.020023999942673575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,16384,0.26292888323465985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,5120,0.015241778559154935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,6144,0.04025244381692674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,2048,0.009526222116417354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,2048,0.02392533255947961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,256,0.021482666333516438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,128,0.004113777644104428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,65536,13.838885837131075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,768,0.005225777626037598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,2560,0.008112000094519721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,2048,0.013097777962684631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,51200,1.6642381880018446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,512,0.007082666787836287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,2560,0.047325332959493004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,8192,0.015793777174419828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,12288,0.12343466281890869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,51200,1.1374390920003254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,16384,0.21210843986935088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,32,0.012737777498033313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,512,0.009878222313192155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,3072,0.019861333900027804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,2560,0.014287999934620328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,32,0.005940444353553984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,51200,0.22709332572089302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,51200,0.10000889168845283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,512,0.004375111311674118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,1536,0.04724355538686117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,51200,0.13650933901468912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,512,0.006855999843941794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,65536,1.5634302563137477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,3072,1.1964346567789714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,1536,0.023338667220539514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,256,0.003779555360476176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,2048,0.00817599975400501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,51200,1.716743999057346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,3584,0.016929777132140267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,7168,0.006895111252864202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,2560,0.032305777072906494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,3584,0.050563553969065346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,51200,0.6867341995239258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,3584,0.011765333513418833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,128,0.00740622232357661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,2560,0.007344889144102733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,65536,0.09955288966496785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,768,0.013519110778967539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,1024,0.007423111134105259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,10240,0.013834666874673633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,256,0.0030053332448005676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,32,0.004965333475006951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,1024,0.0058959999846087555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,1024,0.012109333442317115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,2560,0.011226666470368704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,6144,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,4096,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,5120,0.011266666981909009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,2560,0.02277333372169071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,6144,0.04483022292455038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,3072,0.011557333171367645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,3584,0.14189955923292372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,256,0.003375111147761345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,65536,0.4708399772644043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,3584,0.18124355210198295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,5120,0.03601600064171685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,1536,0.04486666785346138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,512,0.014139556222491793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,12288,0.010127999716334874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,12288,0.07246222098668416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,2560,0.0063777777055899305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,64,0.0053119998839166425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,32,0.018370666437678866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,3584,0.07188889053132799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,3072,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,3072,0.01521511044767168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,256,0.004332444320122401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,7168,0.2650417751736111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,1024,0.0064737776087390045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,51200,1.5201644897460938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,2560,0.007183999651008182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,5120,0.02072888943884108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,4096,0.00850399997499254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,6144,0.04370133413208855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,5120,0.018226666582955253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,5120,0.020491555333137512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,768,0.009604444106419882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,256,0.012176888684431711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,65536,1.232137786017524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,16384,0.009726222190592024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,768,0.004512000001139111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,2560,0.009361777868535783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,32,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,128,0.006002666635645761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,16384,5.963905758327908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,2560,0.04475466741455925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,5120,0.02087466584311591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,5120,0.09098577499389648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,512,0.004305777864323722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,512,0.03671822283003066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,512,0.005629333357016246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,5120,0.02868888775507609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,7168,0.034809778134028115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,2560,0.006002666635645761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,2048,0.02293866707219018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,2560,0.23039200570848253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,1024,0.06161866585413615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,5120,0.03895644346872965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,1536,0.006993778049945831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,1536,0.04396533303790622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,6144,0.006949333681000604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,12288,1.30185423956977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,1024,0.03646399908595615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,5120,0.0177848885456721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,1024,0.007711110843552484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,2560,0.013253333667914072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,51200,1.5142684512668187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,8192,0.01309333327743742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,65536,0.12740800115797254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,1024,0.00416355538699362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,1536,0.009916444619496664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,512,0.010974222587214576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,16384,0.1407111088434855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,12288,0.12923288345336914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,8192,0.27227555380927193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,10240,0.08309955729378594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,128,0.027031110392676458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,3584,0.088100446595086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,768,0.005571555760171678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,256,0.013105777402718862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,32,0.015100444356600443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,768,0.05127733283572727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,10240,0.02869066596031189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,5120,0.03264355659484863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,10240,0.8183706601460775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,64,0.04012177719010247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,1024,0.009089777866999308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,3072,0.05131555596987406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,51200,11.657500372992622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,128,0.015994666351212394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,10240,0.04862133330769009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,2560,0.01443733274936676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,16384,0.12161244286431207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,6144,0.006341333190600078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,5120,0.0068853331936730286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,5120,0.014328888720936246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,3584,0.07865333557128906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,64,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,2560,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,2048,0.01575377749072181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,1536,0.07588266664081149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,5120,0.008360889222886827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,1024,0.06119822131262886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,2560,0.01577333278126187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,8192,0.026025777061780293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,4096,0.020291555258962844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,1536,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,7168,0.027048889133665297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,64,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,768,0.005688888745175467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,1536,0.006478222293986215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,2560,0.05990755558013916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,12288,0.05631733602947659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,8192,0.21919111410776773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,65536,0.07135288582907783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,7168,0.040447112586763166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,128,0.007735999921957652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,32,0.18405421574910483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,2560,0.18545599778493246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,1536,0.004864000197913912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,256,0.01400888959566752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,3072,0.9763911565144857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,3072,0.013648889131016202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,65536,0.15359200371636286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,8192,0.15840000576443142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,12288,0.013560888667901358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,1024,0.007356444166766272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,1536,0.016407999727461074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,4096,0.035804443889194064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,256,0.003855111284388436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,5120,0.009280888570679558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,2048,0.01035733355416192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,32,0.0158942225906584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,2048,0.055516444974475436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,8192,0.026761778526835974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,64,0.005179555465777715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,7168,0.008114666574531132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,1024,0.007094221810499827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,6144,0.023056889573733013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,3072,0.0397093329164717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,3072,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,65536,0.07535466882917616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,768,0.05103111267089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,65536,0.13191466861301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,128,0.007057777709431118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,1024,0.05061866508589851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,2048,0.017179555363125272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,6144,0.17614844110276964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,12288,0.01923822197649214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,6144,0.05045244428846571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,128,0.010679110884666443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,5120,0.008782222039169734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,1536,0.024513777759340074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,16384,1.4342595206366644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,16384,0.029126223590638902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,512,0.009492444495360056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,3584,0.00923733330435223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,64,0.006597333484225803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,10240,0.017530666457282174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,2560,0.008805333740181392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,2560,0.012544000314341651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,12288,0.02256800068749322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,6144,0.011946666571829053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,8192,0.01608444419172075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,768,0.16557599438561335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,4096,0.02607288956642151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,1024,0.013103110922707452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,7168,0.2946755621168348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,2560,0.06851289007398817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,10240,0.6131564246283637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,768,0.011572444604502784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,51200,0.08708710802925958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,3584,0.06737066639794244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,51200,1.3153520160251193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,768,0.0038559999730851916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,3584,0.03035822179582384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,2560,0.030466665824254353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,12288,0.15462666087680393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,6144,0.013024888932704926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,256,0.010763555765151978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,16384,0.040997332996792264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,7168,0.0136835558546914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,1536,0.05623910824457804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,3584,0.007821333077218797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,5120,0.011561777856614856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,1536,0.21621244483523896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,256,0.023557333482636347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,768,0.02732444471783108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,3584,0.013225778109497495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,1024,0.00637155564294921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,10240,0.019696000549528334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,2048,0.016336000627941556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,64,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,1536,0.01882933411333296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,256,0.0248346659872267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,4096,0.00830488900343577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,64,0.005286222116814719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,10240,0.15775644779205322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,5120,0.006849777781301075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,16384,0.4503359794616699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,12288,0.20246756076812744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,51200,8.197769165039062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,10240,0.025647999511824712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,2560,0.010824888944625854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,8192,0.021048888564109802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,4096,2.878793716430664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,1536,0.020595555504163105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,256,0.002973333415057924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,51200,1.233359972635905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,10240,0.014159111513031853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,2048,0.03708800011210971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,10240,0.09762666622797649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,6144,0.014685332775115967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,3072,0.016018667154841952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,65536,0.0746284458372328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,12288,0.9198889202541776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,3072,0.014614222778214348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,4096,0.0750675532552931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,65536,0.2390355534023709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,4096,0.15695377190907797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,1536,0.025711999999152288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,5120,0.016113777955373127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,1536,0.03849600089920892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,16384,0.03160888950030009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,32,0.004209777961174647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,1024,0.011371555427710215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,12288,0.2678195635477702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,6144,0.030828442838456895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,2560,0.007510222494602203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,7168,0.5312142372131348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,65536,0.0988231102625529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,12288,0.027662222584088642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,128,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,768,0.28039643499586314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,5120,0.01222222215599484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,5120,0.009176889227496253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,64,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,7168,0.019709333777427673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,8192,0.017451556192504037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,7168,0.0691866676012675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,4096,0.0165928883685006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,768,0.06924266947640313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,7168,0.07841866546207003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,768,0.006351999938488007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,8192,0.5002986590067545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,12288,0.038514667087131076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,12288,3.733582390679253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,1024,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,1024,0.014190221826235453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,2560,0.024785778588718835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,256,0.020079111059506733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,768,0.07118311193254259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,3584,0.015073777900801765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,256,0.003080000066094928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,10240,0.08781866894827949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,65536,15.252204047309029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,51200,0.07756266991297404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,5120,0.03442755672666762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,51200,0.08699644274181789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,512,0.004640000147951974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,32,0.12797421879238552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,8192,0.015256888336605497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,3072,0.008055999875068665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,16384,0.45809422598944766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,3072,0.009088888764381409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,768,0.012724444270133972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,16384,0.654711988237169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,1024,0.04888800117704603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,768,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,768,0.00423555572827657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,16384,0.036072889963785805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,10240,0.275544007619222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,32,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,1024,0.009404444032245213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,4096,0.410808006922404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,65536,0.2810737821790907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,2560,0.015977778368526034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,8192,0.019126221537590027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,32,0.007607111500369177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,10240,0.03327288892534044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,16384,0.38826576868693036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,10240,0.014922666880819531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,51200,29.00088840060764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,2560,0.060079998440212674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,12288,0.7539777755737305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,16384,1.5729804568820531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,1024,0.047784000635147095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,1536,0.013696889082590738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,1536,0.006033777776691649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,8192,0.016394666499561734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,1536,0.6591546800401475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,768,0.007256000406212277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,5120,0.011055111057228513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,8192,0.191302220026652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,5120,0.16961688465542266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,12288,0.24421599176194933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,2560,0.03166488806406657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,512,0.006116444451941385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,768,0.003887999802827835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,8192,0.12017867300245498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,12288,0.030512889226277668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,16384,0.914921760559082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,512,0.022361778550677832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,16384,0.012459555433856117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,6144,0.012783110969596438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,65536,3.4725831349690757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,5120,0.014003554979960123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,65536,0.43243998951382107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,256,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,65536,0.05850133630964491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,256,0.008246222303973304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,8192,0.0285173323419359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,16384,6.704626719156901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,6144,0.09142488903469509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,51200,0.7932275666130914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,128,0.17401956187354195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,32,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,8192,0.02576444380813175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,8192,0.21163200007544625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,256,0.020023999942673575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,65536,0.15176088280147976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,2048,0.0052782222628593445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,16384,0.4928995768229167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,16384,0.025075554847717285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,1024,0.006600889066855113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,3584,0.01864088906182183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,2048,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,128,0.009381333159075843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,2048,0.013642667068375481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,32,0.1664337714513143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,7168,0.04671733246909248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,2048,0.010515555739402771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,7168,0.011575111084514193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,4096,0.01107022249036365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,256,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,2048,0.08204533656438191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,8192,0.04507911205291748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,64,0.012917333179050021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,1536,0.009813333551088968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,3584,0.009771555662155151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,1024,0.03882222374280294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,1536,0.344105773501926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,8192,0.040834665298461914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,4096,0.2516106764475505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,7168,0.0655404461754693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,12288,0.2823928991953532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,4096,0.027300443914201524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,1024,0.007428444094128079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,1536,0.007028444773621029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,3584,0.19697954919603136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,2560,0.006800000038411882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,2560,0.006129777679840724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,1024,0.005524444497293896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,7168,0.04051288962364197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,5120,0.0805822213490804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,51200,0.19991555478837755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,512,0.005972444597217772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,8192,0.07975377639134724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,1536,0.012868444124857584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,51200,0.008417777717113495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,16384,0.2541191048092312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,6144,0.15596800380282932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,32,0.0033573332346147964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,5120,0.01719377769364251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,3584,0.016760889026853774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,2048,0.06728621986177233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,12288,0.039503110779656306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,4096,0.6725200017293295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,256,0.003338666632771492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,4096,0.2926870981852214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,1536,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,64,0.022135999467637803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,12288,0.08594666586981879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,768,0.028324445088704426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,16384,0.5831528769599067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,51200,0.37809867329067653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,256,0.011552888486120435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,1536,0.006753777878151999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,5120,0.008900444540712569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,1536,0.02532800038655599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,3584,0.13735911581251356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,64,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,128,0.02016533414522807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,128,0.0026977778308921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,2560,0.04268977708286709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,128,0.003012444410059187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,51200,3.065413369072808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,512,0.004596444467703502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,256,0.011003555523024665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,2560,0.08779644303851658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,2560,0.14447555277082655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,256,0.004453333301676644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,16384,0.4708915816413031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,7168,0.01367911116944419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,2048,0.006983111302057902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,12288,1.29713069068061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,6144,0.010250666903124915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,1536,0.016620443926917184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,10240,0.10644977622561985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,6144,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,2560,0.07530399825837877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,1024,0.019316444794336956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,128,0.006762666834725275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,3584,0.054679112301932446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,5120,0.011439999772442712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,32,0.0026346666531430352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,768,0.013305777476893531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,10240,0.1525884469350179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,12288,0.05689777930577596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,65536,9.863416035970053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,1024,0.012919999659061432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,2560,0.01016088906261656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,6144,0.04197066691186693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,3072,0.010581333604123859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,2048,0.0658408866988288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,3584,0.009752000371615091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,64,0.010384000009960597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,12288,0.04737955662939283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,12288,0.00628266649113761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,1024,0.013179555535316467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,64,0.014722666806644864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,4096,0.056038220723470054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,5120,0.47574933369954425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,3584,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,768,0.005959110955397288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,7168,0.11805066797468398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,16384,0.5302524566650391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,32,0.0037235555549462638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,2048,0.008000888758235509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,2560,0.03660444418589274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,51200,1.7227289411756728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,8192,0.0481697784529792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,1024,0.007968000239796108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,7168,0.06402933597564697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,128,0.03788977861404419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,3072,0.014113777213626437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,768,0.011855999628702799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,10240,0.12212088372972275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,16384,0.10384177499347264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,768,0.009708444277445475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,8192,2.274113761054145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,16384,0.41492178705003524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,4096,0.012103999654452005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,7168,1.5218640433417425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,3072,0.26994487974378795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,128,0.013131555583741931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,7168,0.015571556157535978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,1536,0.0162791113058726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,64,0.003107555624511507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,64,0.005662222289376789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,64,0.003921777837806278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,1536,0.06923288769192167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,512,0.1609813372294108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,2048,0.007542221910423703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,3584,0.0811306635538737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,32,0.002522666628162066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,8192,0.03109244505564372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,768,0.019052444232834708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,5120,0.06734310918384127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,5120,1.4878738191392686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,4096,0.025136000580257837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,2560,0.02797777785195245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,1536,0.05456533365779453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,7168,0.03597866826587253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,6144,0.4421644475724962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,10240,0.01648000048266517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,6144,0.006287999865081575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,2048,0.1785902182261149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,256,0.004661333229806689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,1536,0.009576000273227692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,64,0.00509688879052798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,4096,0.5368231137593588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,2560,0.034154666794670954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,51200,0.17795022328694662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,2048,0.00831822223133511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,256,0.0962302221192254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,512,0.036970665057500206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,256,0.021738666627142165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,65536,1.5123102400037978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,128,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,2048,0.008226667013433244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,256,0.004818666726350784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,1536,0.36697689692179364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,128,0.0062871111763848206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,1024,0.022443556123309668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,7168,0.017616889543003507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,8192,0.0313022236029307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,3072,0.009045333498054082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,1536,0.006182222316662471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,10240,0.0270115551021364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,2048,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,65536,0.21399999989403617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,6144,0.01610933409796821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,2560,0.0655075576570299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,1536,0.007099555598364935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,2560,0.010962666736708747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,1536,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,2048,0.04762844575775994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,5120,0.009694221946928237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,3584,0.01587022178702884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,10240,0.013651555611027611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,32,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,3072,0.009283555878533257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,7168,0.0380311111609141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,768,0.010248888697889116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,7168,0.05677511294682821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,1024,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,4096,0.010065777434243096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,64,0.007647111184067196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,2048,0.0074515557951397365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,4096,0.009783999787436591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,1536,0.025627555118666753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,10240,0.48090044657389325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,32,0.0075697774688402815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,512,0.009450666606426239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,51200,0.08251377609041002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,51200,1.011950174967448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,10240,0.019526221685939364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,2560,0.07363466421763103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,128,0.0035297779573334586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,5120,0.03107822272512648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,64,0.006078222145636876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,128,0.0026951111439201566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,12288,0.12763910823398167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,16384,0.007941332956155142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,2560,0.0074026667409473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,7168,0.03177244464556376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,65536,0.20457155174679229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,2048,0.0324577788511912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,128,0.005929777605666055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,128,0.003891555385457145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,1024,0.07846399810579088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,3584,0.020972443951500788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,1536,0.0696622199482388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,2560,0.03934399949179755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,2048,0.03398844599723816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,7168,0.017704889178276062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,12288,0.02513155506716834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,1024,0.011804444094498953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,8192,0.12273689111073811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,51200,2.3275519476996527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,768,0.008397333323955536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,6144,0.02525955604182349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,3584,0.010363555616802638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,2560,0.01643200053109063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,32,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,3584,0.03976088762283325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,65536,0.1579564412434896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,51200,0.10440088642968072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,64,0.002671111168132888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,2048,0.011142222417725457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,32,0.015604444675975375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,16384,0.06665777497821383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,8192,0.056364443567064076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,5120,0.01219377749496036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,1536,0.01016622202263938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,2048,0.016938666502634685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,12288,0.0980222225189209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,8192,0.03499911228815714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,512,0.0038240001433425476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,3584,0.022247999906539917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,5120,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,2048,0.00740444411834081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,8192,0.060277336173587374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,6144,0.008260444634490544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,8192,0.061236441135406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,6144,0.24835377269321016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,256,0.00794311116139094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,7168,0.01941244469748603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,3072,0.01962222158908844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,6144,4.921994527180989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,2048,0.011132444772455426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,8192,0.013718222578366598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,5120,0.032207998964521624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,65536,0.5451119740804037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,8192,0.02492444382773505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,1536,0.49218045340643984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,65536,0.1477831072277493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,6144,0.03976000017589993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,16384,0.021895999709765118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,128,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,51200,0.19755289289686415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,8192,0.20496800210740831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,65536,0.03565244542227851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,64,0.01019377758105596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,2048,0.009144888983832465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,16384,0.05273333191871643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,16384,0.024876443876160517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,768,0.004318222403526306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,128,0.0028568889117903183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,2560,0.019949333535300363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,3584,0.00796622203456031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,256,0.006913777854707506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,12288,0.03716533382733663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,3072,0.10371644629372491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,16384,0.06837422317928739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,4096,0.018996444013383653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,2048,0.01219288922018475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,128,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,256,0.007150222029950883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,768,0.005968889014588461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,2048,0.6575644281175401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,1024,0.016399110356966656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,16384,0.4929493268330892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,16384,0.022248889009157818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,4096,0.24667644500732422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,10240,0.2563084496392144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,1024,1.7300151189168294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,3072,0.012624888784355588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,3584,0.07537155681186251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,1536,0.05916355715857612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,10240,0.0158897770775689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,7168,0.3017013337877062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,2048,0.22306754853990343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,5120,0.39607734150356716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,256,0.0032711111836963226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,768,0.012200888660218982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,5120,0.02291733357641432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,2560,0.08103289206822713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,51200,0.05800177653630575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,8192,0.01909955508179135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,128,0.0064088888466358185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,5120,0.010055111514197456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,3584,0.13023466534084743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,1536,0.011716444459226398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,256,0.005274666680230035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,16384,0.03031022349993388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,2560,0.061687999301486544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,65536,0.27058932516309947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,5120,0.38003378444247776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,12288,0.069577780034807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,12288,0.02132355504565769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,512,0.03876533442073398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,3584,0.019183999962276883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,128,0.009959110783206092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,4096,0.027975112199783325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,8192,0.6669244236416287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,8192,0.04511022236612108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,16384,0.08966666460037231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,12288,4.3527954949273004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,6144,1.9725733862982855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,65536,13.633077833387587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,7168,0.2196800046496921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,4096,0.013942221800486246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,10240,0.04336533281538221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,128,0.004543111142185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,4096,0.0374257763226827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,65536,0.07273244195514254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,3072,0.061262223455641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,4096,0.016371554798550077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,16384,1.2148204379611545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,3584,0.023465777436892193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,4096,0.10798311233520508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,8192,0.05647110939025879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,12288,0.014797333214018079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,4096,0.09442222118377686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,256,0.004620444443490771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,51200,0.01294311136007309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,8192,0.03295288814438714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,7168,0.058693336115943066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,16384,0.2970835632748074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,128,0.0063271112740039825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,10240,0.590520011054145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,256,0.006862222320503659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,32,0.003105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,2560,0.08076355854670207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,8192,0.024126221736272175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,3584,0.24991467263963488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,1536,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,16384,0.23329332139756945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,128,0.0028977776981062363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,768,0.16772711277008057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,128,0.1262764400906033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,3584,0.013255111045307584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,51200,0.07010310888290405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,32,0.008906666603353288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,4096,0.03712177938885159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,3584,0.010460444622569613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,64,0.005072888814740711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,65536,0.12539022498660618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,1024,0.003996444245179494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,12288,0.04743555519315931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,12288,0.042224887344572276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,768,0.010857777463065254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,512,0.010747555229398938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,2048,0.011576889289749993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,128,0.016884444488419425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,64,0.008090666598743862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,4096,0.11647378073798285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,512,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,4096,0.029571556382709082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,6144,0.017955554856194388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,5120,0.013889777991506787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,256,0.02866311205757989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,32,0.006184000107977126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,51200,0.08760799964269002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,12288,0.7029502126905652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,128,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,2048,0.012239110966523489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,10240,0.01944266590807173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,768,0.005741333381997214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,12288,0.019863999552196927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,12288,0.018812444474962022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,8192,0.012999110751681857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,51200,0.4272142251332601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,512,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,2560,0.008751110898123847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,8192,0.06892800331115723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,7168,0.014366222752465142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,2048,0.0068915556702348925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,8192,0.07159288724263509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,32,0.005666666560702854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,64,0.0028035555862718155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,51200,0.15697067313724095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,10240,0.01610933409796821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,128,0.004310222135649787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,128,0.01682044400109185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,1024,0.006186666587988536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,128,0.009360888765917884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,3584,0.015427556302812366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,768,0.005200888961553574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,128,0.013487999637921652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,51200,0.8018302387661405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,2560,0.048708445496029325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,2560,0.1563040018081665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,12288,0.05076711045371162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,5120,0.023992000354660883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,16384,0.23221243752373588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,2560,0.030654221773147583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,1536,0.029930667744742498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,4096,0.07109333409203424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,8192,0.10242222415076362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,128,0.009012444151772393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,256,0.007797333101431529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,256,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,64,0.023021333747439917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,2560,0.009754666851626502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,4096,6.253412458631728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,2048,0.01051111105415556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,16384,0.25407910346984863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,5120,0.020424000091022916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,1024,0.0047324444684717385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,1536,0.11396977636549209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,51200,5.367433759901259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,8192,0.009663110805882348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,128,0.0035146665241983202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,3584,0.9522764417860242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,256,0.010594666832023196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,3584,0.04309689005215963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,512,0.12692622343699136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,256,0.010185778141021729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,128,0.002790222151411904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,10240,0.04452177882194519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,5120,0.020042666130595736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,3072,0.009752888646390703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,3072,0.15021599663628474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,3584,0.040200889110565186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,1024,0.0199946661790212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,10240,0.046874665551715426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,10240,0.015641777051819693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,64,0.002760888801680671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,128,0.00444444434510337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,256,0.019038221902317472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,51200,0.10977600018183391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,2560,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,12288,0.03280177712440491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,2048,1.5762898127237956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,3072,0.0118231111102634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,51200,2.3141600290934243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,8192,0.042543109920289784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,10240,0.01903288894229465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,7168,0.037112000915739275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,16384,7.425798204210069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,64,0.02696711156103346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,16384,2.3795867496066623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,5120,0.015786666009161208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,7168,1.6006266276041667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,2560,0.00847111145655314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,3072,0.23452801174587676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,2560,0.008561777571837107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,3584,0.017879999346203275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,1536,0.009663999908500248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,3072,0.172343995836046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,64,0.0026142222599850762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,256,0.0031600000543726813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,3072,0.007655999726719326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,1024,0.006220444622966979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,4096,0.030064887470669214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,128,0.003181333343187968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,6144,0.010877333581447601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,51200,0.21390755971272787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,5120,0.00869599978129069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,5120,0.21439466211530897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,16384,0.011865778101815118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,768,0.012967999610635968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,2560,0.01092711091041565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,2048,0.0075795559419526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,768,0.05942844682269626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,32,0.0033119999700122406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,512,0.014494222071435718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,3584,0.013266666895813413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,512,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,256,0.004430222428507275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,2560,0.0063768890168931745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,8192,0.5991102324591743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,32,0.005970666805903117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,7168,2.7473360697428384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,4096,0.07119733095169067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,128,0.034119112624062434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,3584,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,768,0.00904266701804267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,4096,0.008191111187140146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,4096,0.019351111518012155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,12288,0.01738844480779436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,2560,0.21799200110965303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,12288,0.006496888895829518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,1024,0.00908799966176351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,4096,0.005933333188295364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,16384,0.11470488707224528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,6144,0.08930577834447224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,65536,0.016159999701711867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,32,0.0025439999169773525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,10240,5.604445139567058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,51200,0.18783556090460884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,768,0.0074942219588491656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,1024,0.006063111126422882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,4096,0.12224000030093723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,2560,0.07366489039527045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,51200,0.008747555315494537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,65536,3.2448719872368703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,128,0.009487110707494948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,16384,7.931919521755642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,1536,0.04706666535801358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,256,0.009709333380063375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,51200,1.4925368626912434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,16384,0.23270755343967012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,1536,0.010583110981517367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,2048,0.013948443863126965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,4096,0.11579999658796523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,4096,0.015689777003394235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,768,0.005759999983840519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,3072,0.13118578328026667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,128,0.0033200000309281876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,1536,0.00998399986161126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,256,0.006744888921578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,512,0.0034497777620951333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,8192,0.02198222279548645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,51200,0.23436178101433647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,10240,0.008028444316652086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,1536,0.0047848886913723415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,5120,0.017873777283562552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,768,0.008400888906584846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,2048,0.10930577913920085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,1024,0.01036977767944336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,6144,0.018457776970333524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,3584,0.06699377960628934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,4096,0.7420293490091959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,65536,0.31210043695237905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,768,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,1024,0.03501866592301263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,12288,0.0965226689974467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,51200,7.780023362901475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,4096,0.015087111128701104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,4096,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,512,0.009726222190592024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,5120,0.027941332923041448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,768,0.01889866590499878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,8192,0.12548444006178114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,4096,0.34917333390977645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,32,0.004197333421972063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,4096,0.08794222275416057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,7168,0.5537537998623318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,3072,0.0105262224872907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,65536,0.12585688961876765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,10240,0.3390035629272461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,2560,0.009022222624884712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,3072,0.013656888571050433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,5120,0.009577777650621202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,64,0.0032106666929192017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,64,0.008861333131790161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,256,0.006575111299753189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,2048,0.017853332890404593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,128,0.013944889108339945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,4096,0.42084089914957684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,4096,1.0132320192125108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,1536,0.07521155807707045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,8192,0.21082754929860434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,12288,0.4003342257605658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,16384,0.02683111031850179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,1536,0.6860684288872613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,51200,0.29781421025594074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,4096,0.048904889159732394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,12288,0.17995999919043648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,2560,0.02264533274703556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,256,0.012038222617573209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,256,0.037668443388409086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,6144,0.16188711590237087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,1024,0.06963733169767591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,12288,0.02886399957868788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,5120,0.009420444567998251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,256,0.012660444610648684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,128,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,512,0.0037048889531029593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,10240,0.051354666550954185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,2048,0.44367289543151855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,2560,0.019129777948061626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,128,0.004457777573002709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,1536,0.01128266668981976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,1536,0.005118222286303838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,256,0.011146666275130378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,65536,0.010841777755154504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,12288,0.017817777064111497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,2048,0.02108088963561588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,6144,0.014935111006100973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,6144,0.016887111796273124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,65536,0.6112657652960883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,2048,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,51200,0.021624000536070928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,256,0.005260444349712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,10240,0.01644000079896715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,8192,0.10147288772794937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,2560,0.009949333137936061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,1536,0.08441955513424343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,7168,0.06913244724273682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,5120,0.04653155472543505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,4096,0.007931555310885111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,3072,0.01720888912677765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,768,0.01146577795346578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,8192,0.02216888964176178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,4096,0.10692533519532944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,65536,4.180204603407118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,10240,0.46927913029988605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,8192,1.1283084021674263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,51200,0.2419688966539171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,768,0.0041573333243529005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,6144,0.02774311105410258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,1024,0.030976888206270006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,1536,0.014202665951516895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,7168,0.2718417909410265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,1024,0.03728444377581278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,12288,0.006314666734801398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,1536,0.010402667025725046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,3072,0.007915555602974361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,7168,0.1920177804099189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,16384,0.03940711087650723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,3072,0.016033778587977093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,51200,0.06902311245600383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,10240,0.16698222690158418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,3072,0.011212444139851464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,3072,0.01145511120557785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,3072,0.38964088757832843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,10240,0.015855999456511605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,5120,0.006804444309737947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,3584,0.020319999919997323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,10240,0.3234684467315674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,12288,0.07924533552593656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,3584,0.022135999467637803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,256,0.015512888630231222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,3072,0.19887288411458334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,7168,0.037235554721620344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,32,0.010246222217877706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,6144,0.016262221667501662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,7168,0.18677688969506157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,4096,0.06472977664735582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,6144,0.02979644470744663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,3072,0.018743111027611625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,256,0.0033128888656695685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,256,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,3072,0.12959199481540257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,3072,0.8265920215182834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,3584,0.09735288884904648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,5120,0.018340445227093168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,7168,0.015792000624868605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,16384,0.025193777349260118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,1024,0.06100800302293566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,1536,0.011222222612963783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,10240,0.017063111066818237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,7168,0.23013599713643393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,5120,0.08094844553205702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,256,0.04658755660057068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,16384,0.04902044600910611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,3072,0.006878222028414409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,512,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,2560,0.017796445224020217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,3584,0.020710221595234342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,16384,0.05035733183224996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,256,0.005669333454635408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,128,0.0059573331640826324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,64,0.0026666666898462507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,8192,0.028258668051825628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,768,0.015088000231319003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,2048,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,5120,0.015415110521846347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,2560,0.02125777800877889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,128,0.0029128889242808023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,12288,0.02572622232966953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,2560,0.007127111156781514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,2048,0.01108088923825158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,16384,3.4485191769070096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,1024,0.024518221616744995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,768,0.015538665983412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,4096,0.010320000350475311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,10240,0.021133333444595337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,64,0.008653333617581261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,10240,0.6782497829861112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,3072,0.009999111294746399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,65536,5.354097578260634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,12288,0.023756444454193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,8192,0.015743111570676167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,768,0.016805332567956712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,6144,0.007511999871995714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,2048,0.011709333293967776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,768,0.05218755536609226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,51200,0.10094400246938069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,1024,0.02440888848569658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,7168,0.11448000537024604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,2048,0.12645510832468668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,512,0.007100444700982835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,1536,0.01644000079896715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,1536,0.008256889051861232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,6144,0.16007644600338405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,4096,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,10240,0.08403733041551377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,256,0.009214222431182861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,10240,0.03801866703563266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,512,0.0033599999215867785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,5120,0.013292444248994192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,51200,2.7160284254286022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,5120,0.02485688858562046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,7168,0.25451376703050405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,10240,0.1344204478793674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,2048,0.011431111229790581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,64,0.006512889017661412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,256,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,1536,0.004808889081080754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,6144,0.10693511035707261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,51200,0.12864800294240317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,2560,0.016392888294325936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,3584,0.11707022455003525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,256,0.02663644486003452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,1024,0.006433777925040986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,10240,0.07682933410008748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,1024,0.004628444297446145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,512,0.02109955582353804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,7168,0.020489777127901714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,65536,4.235791948106554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,7168,0.025998221503363714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,16384,0.029337777031792536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,2560,0.030424889591005113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,768,0.005988444305128521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,3072,0.31511023309495717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,2560,0.007092444433106317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,1024,0.009905777871608734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,1024,0.006320888797442119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,256,0.003087999920050303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,5120,0.21755999988979766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,3584,0.02469066613250309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,16384,0.07022755675845675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,1024,0.009768000079525841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,1536,0.07397422525617811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,4096,0.012444444828563266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,2048,0.03386844529045953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,768,0.01406311160988278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,32,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,768,0.005719111197524601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,12288,0.2621440092722575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,512,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,51200,9.996691385904947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,51200,0.9766427146063911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,65536,0.10858133104112412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,7168,0.9900177849663628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,3584,0.05087022317780388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,512,0.0053022222386466126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,4096,0.009387555221716562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,3584,0.01052977806992001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,768,0.013072888884279462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,8192,0.10112177663379246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,3584,0.017717333303557504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,768,0.011182222101423474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,32,0.027084444959958393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,51200,0.06051644351747301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,128,0.005405333307054307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,1536,0.006787555499209299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,10240,0.02548355526394314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,3584,0.011711111499203576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,10240,0.01463822192615933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,7168,0.018744000130229525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,6144,0.18706933657328287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,12288,0.2396311230129666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,2560,0.007154666715198093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,1024,0.40720356835259336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,6144,0.01349955548842748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,256,0.00527022240890397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,5120,0.010737777584128909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,12288,0.411060439215766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,128,0.023880889018376667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,512,0.004327110946178436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,51200,0.11718932787577312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,7168,0.04021155503061082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,5120,0.022072000635994807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,3584,0.01035377797153261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,256,0.12499288717905681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,64,0.006503110958470239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,2048,0.012895999683274163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,10240,0.022245332598686218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,768,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,2048,0.011367999845080905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,16384,0.03188088867399428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,7168,0.020058666666348774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,10240,0.013519110778967539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,256,0.06705066892835829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,768,0.005602666487296422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,12288,0.02160177793767717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,4096,0.01257155587275823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,2048,0.006667555620272954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,256,0.01700266698996226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,51200,0.052002665069368154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,2560,0.01661155621210734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,768,0.003929777691761653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,5120,0.1021875540415446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,2560,0.00795733349190818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,1024,0.040651556518342756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,2560,0.5205573505825466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,65536,0.3602622350056966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,51200,0.3368319935268826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,12288,0.2728293206956652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,768,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,64,0.012598222328556908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,4096,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,4096,0.1634933286243015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,5120,0.07860800292756823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,6144,0.036622222926881574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,128,0.0027235555979940626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,32,0.005408888889683618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,256,0.04038666685422262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,65536,0.07602755228678386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,12288,0.010977778169843884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,7168,0.015850666496488783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,64,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,2560,0.04684977730115255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,3584,0.04172977805137634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,3584,0.010346666806273991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,512,0.07794133159849378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,64,0.030384888251622517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,10240,0.07236266798443265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,65536,0.24485511249966094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,128,0.004480888860093223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,8192,0.006562666760550604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,16384,3.3725501166449656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,2560,0.01621955633163452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,4096,0.009675555759006077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,1536,0.02615822189384037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,4096,0.20073155562082926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,7168,0.3089333375295003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,256,0.004373333520359463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,2560,0.08948889043596055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,65536,0.8853404257032605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,768,0.00620977787507905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,65536,0.20942044258117676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,2560,0.08520533641179402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,2560,0.9955039554172092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,8192,0.024136889312002394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,512,0.0040702223777771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,7168,0.025452444950739544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,10240,0.023502222365803186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,7168,0.02380977736579047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,128,0.004503111044565837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,6144,0.013105777402718862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,6144,0.4718337588840061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,10240,0.11104089021682739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,6144,0.07447377840677898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,65536,0.6538453102111816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,6144,0.4902044402228461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,12288,0.19008089436425102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,2048,0.034656888908810086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,2048,0.009942221972677443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,2560,0.008796444369686974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,2560,0.007184888753626082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,512,0.003540444291300244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,768,0.007576888634098901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,128,0.0031102223114834893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,768,0.06345333655675252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,256,0.0058906666106647914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,256,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,65536,0.04991555545065138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,1536,0.011896889242861005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,4096,0.013927999469969006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,10240,0.042618665430280894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,512,0.004423111263248655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,128,0.03258311086230808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,5120,0.018988443745507132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,3584,2.2696605258517795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,256,0.02861777941385905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,10240,0.13127999835544163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,3072,0.1298417780134413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,3584,0.022897778285874262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,512,0.005912888795137405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,65536,0.06753866540061103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,12288,0.017873777283562552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,16384,0.03759288787841797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,3072,0.05682933330535889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,65536,0.269432889090644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,12288,0.07751555575264825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,32,0.0074720001882976955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,8192,2.438038296169705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,1024,0.0455022222465939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,32,0.00868000007337994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,4096,0.05647022194332547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,4096,0.06324177980422974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,32,0.01716888944307963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,8192,0.0703528920809428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,5120,0.10277511013878717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,65536,0.13359288374582926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,1024,0.007710222568776872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,7168,0.030943999687830608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,51200,0.010402667025725046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,65536,0.11468622419569229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,2048,0.010557333628336588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,51200,0.062112000253465444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,5120,0.05362577901946174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,8192,0.7485191027323405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,7168,0.03572266631656223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,5120,0.011784888803958893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,8192,0.024705777565638225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,3072,0.010151999692122141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,51200,0.08675377898746067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,16384,0.05456799930996365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,10240,0.09095644288592869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,8192,0.14281066258748373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,64,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,512,0.012679110798570844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,12288,0.29710398779975045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,1536,0.006919999917348226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,51200,1.1484355926513672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,128,0.006477333191368315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,1024,0.011075555450386472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,32,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,4096,0.1012808879216512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,2560,0.5874142116970485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,2560,1.864755630493164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,128,0.0511813329325782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,5120,0.14553511142730713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,16384,0.7372044457329644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,5120,0.07351555426915486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,5120,0.2558151086171468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,4096,0.059989333152770996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,12288,2.191163592868381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,2560,0.06537155310312907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,65536,1.4476586447821722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,2048,1.0665031009250217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,8192,0.0681333343187968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,1536,0.1592506700091892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,4096,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,65536,2.457848866780599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,3584,0.06922222508324517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,51200,0.6999626689487033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,2560,0.016470222009552848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,512,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,6144,0.20519467194875082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,12288,0.012242666549152799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,256,0.007786666353543599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,16384,0.34550044271681046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,51200,1.1867129007975261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,10240,3.6850713094075522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,1536,0.040560000472598605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,2048,0.14744355943467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,512,0.0037813331517908308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,32,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,8192,0.1535324388080173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,2560,0.16493599944644505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,12288,0.17733689149220785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,3584,0.018072888255119324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,4096,0.031437334087159895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,1024,0.010212444596820408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,2048,0.00814400033818351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,8192,0.08399822314580281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,12288,0.0418524444103241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,32,0.003100444459252887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,2048,0.10083644257651435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,1024,0.0069439998931354946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,256,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,8192,0.011635555161370171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,3584,0.008852444589138031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,5120,0.5396089024013943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,768,0.007746666669845581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,65536,0.0934151079919603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,7168,0.3450604544745551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,1536,0.0169804443915685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,2048,0.5800071292453343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,2048,0.006609777609507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,1024,0.01056977775361803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,768,0.007601777712504069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,768,0.009200000100665623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,64,0.013270222478442721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,65536,0.16575911309983996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,51200,0.3201991187201606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,51200,0.02340000040001339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,64,0.002514666567246119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,128,0.005471999860472149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,1024,0.0163137780295478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,8192,0.23522488276163736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,5120,0.09672711292902629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,1536,0.22681156794230142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,8192,0.061441779136657715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,3072,0.056961774826049805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,7168,0.056777775287628174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,5120,0.06715022193060981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,2560,0.00758755538198683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,1536,0.008541333178679148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,256,0.009296000003814697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,1024,0.004302222281694412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,6144,0.014189332723617554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,8192,0.26282578044467503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,10240,0.25008710225423175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,2560,0.0069422221018208405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,256,0.009591110878520543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,768,0.011896889242861005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,16384,0.6183786922030979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,2048,0.04182933436499702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,3584,0.007118222614129384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,6144,0.14684177769554987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,65536,0.04100622071160211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,8192,0.048166222042507596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,768,0.008077333370844523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,1536,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,3584,0.025789333714379206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,1024,0.012670222255918713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,12288,0.7605359819200305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,2560,0.03324177861213684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,256,0.004655111167165968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,16384,1.3961893717447917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,12288,0.011541333463456897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,65536,0.16628889242808023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,3584,0.04318488968743218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,8192,0.013926222920417786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,12288,6.150775909423828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,3072,0.010562666588359408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,65536,0.5924400223626031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,5120,0.15524089336395264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,1536,0.05343733231226603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,5120,0.03756355577045017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,16384,0.10453333457310994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,2560,0.0691208905643887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,5120,0.11534933249155681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,6144,0.5672764248318142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,768,0.005008888741334279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,3584,0.018360000517633226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,6144,0.29537423451741535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,16384,0.03850577937232123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,8192,0.159588442908393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,64,0.0028248888750871024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,3584,0.01759022143152025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,2560,0.0554328891966078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,51200,1.4253351423475478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,512,0.009159111314349705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,12288,1.4636035495334203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,5120,0.13382932874891493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,10240,0.14121245013342962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,8192,0.014284445179833306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,4096,0.02204622162712945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,2048,0.008373333348168267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,16384,0.6054124302334255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,65536,14.4076419406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,32,0.012169777519173093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,10240,0.3382408883836534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,2560,0.010643555886215635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,128,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,8192,0.014873777826627096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,8192,0.008343111309740279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,16384,0.11826222472720677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,3584,0.012729778057999082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,65536,0.4615946875678168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,6144,0.050663110282686025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,2560,0.02458933326933119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,3584,0.143321779039171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,256,0.004838222016890843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,3584,0.015749333633316886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,512,0.005129777722888523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,16384,1.124962700737847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,128,0.005981333139869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,16384,0.04682044519318474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,1024,0.0069546666410234236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,512,0.007993777592976889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,8192,0.00907911111911138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,8192,0.04509688748253716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,4096,0.00777599960565567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,128,0.003903111235962974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,65536,0.08617511060502793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,64,0.00721866637468338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,7168,0.008199999729792276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,1536,0.008976000050703684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,1024,0.10082755486170451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,65536,0.5303902096218532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,2560,0.014448000325096978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,3584,0.26077156596713597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,2048,0.00832088871134652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,12288,0.5075404379102919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,512,0.03715733355946011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,3584,0.02994666828049554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,768,0.37529510921902126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,256,0.0047386665311124595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,5120,0.017183999220530193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,7168,0.008463111188676622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,10240,0.018603555030292936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,12288,0.02700977855258518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,6144,0.03451555636194017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,2048,0.016283555163277518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,1536,0.015006222658687167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,1024,0.005850666513045629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,3584,0.00979022185007731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,5120,0.17520799901750353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,65536,3.737394544813368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,2048,0.02350133326318529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,51200,0.5248248842027452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,2560,0.9322284062703451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,2048,0.00794044468137953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,2560,0.011707555916574268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,512,0.09073511097166274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,6144,0.25147467189364964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,2048,0.10596799850463867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,12288,0.021743110484547083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,256,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,12288,0.018415111634466384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,3584,0.006616888774765863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,4096,0.024123556084103052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,32,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,7168,4.829673767089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,1024,0.07472266753514607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,2560,0.008998221821255153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,2048,0.03886133432388306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,3072,0.1217377715640598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,8192,0.05628710985183716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,1024,0.004043555508057277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,512,0.009526222116417354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,512,0.055296887954076133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,6144,0.45508442984686953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,51200,2.6185315450032554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,4096,0.023853333459960088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,4096,0.09938666555616592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,12288,0.046718223227394946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,8192,0.2880320019192166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,2560,0.03843288951449924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,10240,0.01944533321592543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,10240,0.07811022467083402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,1024,0.030624889665179785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,7168,0.013247999880048962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,256,0.003162666741344664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,51200,0.05328355564011467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,2048,0.025961778230137292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,256,0.014920888675583733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,16384,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,12288,0.4281884564293756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,12288,3.068148507012261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,128,0.003129777808984121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,5120,1.1196275287204318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,51200,0.2933991220262316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,5120,0.01940088967482249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,512,0.0052577778697013855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,2560,0.13562488555908203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,512,0.005166222237878376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,65536,2.640591091579861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,12288,0.03053155541419983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,256,0.005695111221737332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,2560,0.07447733481725057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,3584,0.02221155497762892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,128,0.0041706665522522396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,10240,0.044779555665122144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,10240,0.016537777251667447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,3072,0.011235555840863122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,12288,2.161017735799154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,65536,0.051033778323067554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,2048,0.025235555238193933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,6144,0.020616888999938965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,65536,0.03231466809908549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,7168,0.02939288814862569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,2560,0.02161066730817159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,256,0.017424000634087455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,10240,0.03231733375125461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,16384,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,256,0.004719999929269155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,16384,0.09915644592709011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,7168,0.010809777511490716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,51200,0.7903377744886609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,4096,0.07527733511394925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,256,0.02474400069978502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,16384,0.04248266749911838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,4096,0.1440240012274848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,3072,0.1514835490120782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,5120,0.12002578046586777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,4096,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,1024,0.00848711116446389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,8192,0.0221368885702557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,2048,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,6144,0.6080942153930664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,1536,0.010509333676762052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,2048,0.057824002371893994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,3584,0.6492088635762533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,32,0.0029644444584846497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,10240,0.03968355390760634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,51200,0.43763822979397243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,6144,0.21071555879380968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,3584,0.1976026693979899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,256,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,6144,0.015208888385030957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,8192,0.012040889097584618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,3072,0.042902221282323204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,7168,0.020482665962643094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,1024,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,51200,0.08254755867852105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,65536,0.03301066822475857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,128,0.003354666754603386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,1024,0.021511110994550917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,128,0.0047253333032131195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,12288,0.059225777784983315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,10240,1.2007937961154513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,12288,1.1929057439168294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,768,0.006963555183675554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,3072,0.023337778117921617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,128,0.006498666687144174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,10240,0.0981946653789944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,10240,0.1764897770351834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,1536,0.06400089131461249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,64,0.0031155554784668815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,1024,0.009321778184837764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,256,0.015951999359660678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,3072,0.026126222478018865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,1024,0.005977777971161737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,16384,1.2566969129774306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,65536,1.2320035298665364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,65536,0.2206568850411309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,10240,0.016346666547987197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,5120,0.014348444011476306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,12288,0.055330667230818004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,12288,0.053426666392220386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,10240,0.016369778248998854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,2560,0.4471413294474284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,12288,0.1645146608352661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,4096,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,10240,0.04147911071777344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,768,0.013921777407328287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,12288,0.05079999897215101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,3584,0.19683467017279732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,3072,0.2093964417775472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,51200,0.13824533091651067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,16384,2.933601803249783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,2560,0.30455732345581055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,51200,0.07190133465660943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,10240,0.34188887808058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,65536,0.07092089123196073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,1536,0.01882222294807434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,10240,0.22546754942999947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,128,0.006839111033413146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,512,0.006907555378145642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,64,0.004650666481918759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,768,0.020894222789340552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,7168,0.24281067318386504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,32,0.1296817726559109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,2048,0.017269333203633625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,8192,0.3985191186269124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,8192,0.077729774845971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,2048,0.03731911049948798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,4096,0.009816889133718278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,2048,0.017941332525677152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,128,0.013635555903116861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,32,0.0028293333533737394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,16384,0.4702311091952854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,65536,0.7914506594340006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,7168,0.10762577586703831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,51200,0.2670977645450168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,256,0.06596977843178643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,7168,0.08720799949434067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,2048,0.6693040000067817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,2560,0.09893688890669082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,3584,0.011997333003415002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,1536,0.0277804434299469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,65536,0.34536178906758624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,128,0.007072000039948358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,10240,0.026055110825432673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,4096,0.021168889270888433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,16384,0.8707280158996582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,768,0.00535111129283905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,1536,0.00980000032318963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,3072,0.03514488869243198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,8192,0.2104862266116672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,8192,0.03157866663402981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,3584,0.04045244389110141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,16384,1.2256284289889865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,512,0.07176533010270861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,51200,0.5645564397176107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,2048,0.058600889311896436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,3072,0.0492524438434177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,64,0.07728266716003418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,32,0.005587555468082428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,10240,0.14271556006537542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,3072,0.008480889101823172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,2048,0.007127111156781514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,3072,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,3584,0.013199110825856527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,51200,0.03538044293721517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,8192,0.04615288972854614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,1536,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,3584,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,256,0.011352000137170156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,128,0.015951111912727356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,2048,0.01423288881778717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,1536,0.03359466791152954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,768,0.02381155557102627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,7168,0.011487111449241638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,16384,0.07414755556318495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,2560,0.27040089501274955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,7168,0.019175110591782462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,512,0.061054223113589816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,768,0.0037288889288902283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,4096,0.01259200026591619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,3072,0.006086222413513396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,1024,0.004693333473470476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,4096,0.030765334765116375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,10240,0.17003022299872506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,16384,0.018951111369662814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,6144,0.07789155509736803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,128,0.033470223347345986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,3072,0.011239999698268043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,1024,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,51200,0.05315377646022373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,256,0.05758400095833672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,7168,0.35342221789889866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,128,0.0047848886913723415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,51200,1.5663777457343206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,3584,0.03480444351832072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,64,0.008671110702885522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,512,0.005886222339338726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,128,0.0032044444233179092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,2560,0.013831111292044321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,768,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,2048,0.0961902207798428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,51200,0.5960444344414605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,5120,0.009864889085292816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,10240,0.016511110795868766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,65536,6.133057064480251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,65536,1.125581317477756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,3072,0.008117333054542542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,65536,0.4087031152513292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,6144,0.024580443898836773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,512,0.027521777484152053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,256,0.003269333392381668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,51200,0.06981777482562594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,1536,0.5595351325141059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,12288,0.3661804464128282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,3584,0.049689776367611356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,1024,0.010182222558392419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,3072,0.012086222569147745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,768,0.006258666515350342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,6144,0.007879999776681265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,256,0.005149333427349727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,2048,0.12597510549757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,2560,0.011143111520343356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,3584,0.09740355279710557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,65536,0.2936275535159641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,16384,0.22336888313293457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,10240,0.12318844265407985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,16384,0.05605955256356133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,51200,0.8760515318976508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,512,0.004596444467703502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,12288,0.052367998494042285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,32,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,768,0.008944888909657797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,5120,0.022239110536045496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,8192,0.052052444881863065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,32,0.00305600009030766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,4096,0.008811555802822113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,7168,0.00907555553648207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,2048,0.005400889035728242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,16384,0.03864622116088867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,7168,0.30292977227105033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,3072,0.00978577799267239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,51200,0.15096710787879095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,2048,0.010678222609890832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,10240,0.018365333477656048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,7168,0.014783110883500842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,2048,0.006794666664467917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,4096,0.008600888980759515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,512,0.03944977786805894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,10240,0.01960177719593048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,256,0.01236977759334776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,3584,0.8544266488817004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,16384,0.6467875374688042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,32,0.003035555490189128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,64,0.005932444499598608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,2048,0.05188799897829691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,512,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,256,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,1536,5.840713077121311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,3584,0.007328888608349695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,128,0.007103111180994246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,6144,0.00980444418059455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,65536,0.05481422278616163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,2048,0.03259733319282532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,8192,0.2801617781321208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,3072,0.012541333006487953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,768,0.033668445216284856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,256,0.010689777632554373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,5120,0.009357333183288574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,512,0.003640000190999773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,8192,2.5986542171902127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,2560,0.01146044416560067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,7168,0.33398845460679794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,65536,0.5808044539557563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,32,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,2048,0.006540444576077991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,10240,0.6815199851989746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,512,0.0058017778727743365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,768,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,3584,0.012409778104888068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,7168,0.007836444510353936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,1024,0.014697778556081982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,6144,0.012356444365448423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,6144,3.0618567996554904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,65536,0.6001440154181586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,51200,0.11608444319831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,10240,0.1842942237854004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,8192,0.014902222487661572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,4096,0.034797334008746676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,1024,0.0063271112740039825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,16384,0.009857777920034196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,12288,0.18488000498877633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,51200,0.4348133405049642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,768,0.004408888932731417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,1536,0.09240000115500556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,10240,0.06228355566660563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,16384,0.018523555662896898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,10240,0.016917333006858826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,3072,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,2048,1.2907582388983834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,32,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,2560,0.009229333036475712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,16384,0.09568622377183701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,51200,0.2570737732781304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,16384,0.05747466617160373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,1536,0.004741333425045013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,2560,0.010117332968446944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,65536,2.4921838972303605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,512,0.12227111392550999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,7168,24.184834798177082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,16384,0.23507022857666016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,2048,0.007159111400445302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,10240,0.08027288648817274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,4096,0.16719288296169707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,51200,0.3082675668928358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,1024,0.010249777800507015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,3584,0.11714400185479058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,16384,0.01850488947497474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,1536,0.009864889085292816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,65536,0.12880799505445692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,65536,0.07560888926188152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,768,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,256,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,256,0.00905955582857132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,16384,0.19404178195529512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,3072,0.018352889352374606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,16384,0.026408889227443274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,8192,0.012461333639091916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,1536,0.07708177963892619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,2560,0.007289778027269576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,5120,0.029248889949586656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,7168,0.02404888967672984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,8192,0.0477111107773251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,12288,4.246358235677083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,65536,0.5641484260559082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,2048,0.007875555919276344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,32,0.002653333254986339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,65536,9.980870564778646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,1024,0.07120355632570055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,7168,0.020833777056799993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,3584,0.010617777705192566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,1024,0.015288000305493673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,32,0.006937777830494775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,7168,0.006510222123728857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,1536,0.016012445092201233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,3072,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,5120,0.10780711306465997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,2048,0.5304497612847222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,5120,0.010099555883142684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,3584,0.036377777655919395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,65536,0.3015955554114448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,2048,0.012319111161761813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,1536,0.5665929052564833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,12288,0.046145776907602944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,10240,0.39026665687561035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,3584,0.009710222482681274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,256,0.032624890406926475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,768,0.028699556986490887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,8192,0.7253528700934516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,51200,0.9213280147976346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,65536,0.3265591197543674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,12288,0.03108711043993632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,65536,0.240410672293769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,256,0.012395555774370829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,8192,0.01829600003030565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,3072,0.01204622205760744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,1536,0.004621333132187526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,5120,0.016708445217874315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,16384,0.06018310785293579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,8192,0.04721066686842176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,1536,0.09469955497317845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,4096,0.09450311130947536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,2048,0.04073777794837952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,2048,0.15179288387298584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,7168,0.03156088789304098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,1536,0.034141331911087036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,32,0.005055111315515307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,6144,0.020990222692489624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,3072,0.1708808872434828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,65536,0.23094844818115234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,5120,0.018896889355447557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,8192,0.15646666950649685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,128,0.014460444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,128,0.013724444641007317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,65536,0.21183021863301596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,51200,1.6773156060112848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,5120,0.0070053330726093715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,16384,0.08349866337246364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,8192,0.018107554978794523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,1536,0.015167999598715039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,16384,0.025855110751258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,8192,0.04439111219512092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,2048,0.07262400123808119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,2560,0.00831911133395301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,5120,0.33245333035786945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,4096,0.026494221554862127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,16384,9.632975260416666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,2048,0.009639110830095079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,3072,0.04699022240108914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,4096,0.16104177633921304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,1536,0.015224888920783997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,6144,0.03200799889034695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,3072,0.008232889076073965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,1024,0.022846221923828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,16384,0.3993351194593642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,7168,0.18781422244177925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,5120,0.007313778003056844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,51200,0.8740960227118598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,12288,0.21354399787055123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,128,0.0028044444819291434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,3584,0.009723555710580613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,65536,0.3651386631859674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,51200,0.6897537973192004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,1536,0.014862222803963555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,10240,0.3764897717369928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,51200,0.11107644769880508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,1024,0.013010666602187686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,256,0.02110044492615594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,2048,0.007961778177155389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,64,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,2560,0.012468444804350534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,2560,0.34609601232740617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,8192,2.3838462829589844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,1536,0.02238133384121789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,4096,0.2426755428314209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,1024,0.0063742221229606206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,3072,0.008650666309727563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,65536,0.0700062248441908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,51200,0.05739821990331014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,6144,5.232014973958333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,128,0.0035377778112888336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,6144,0.01700266698996226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,10240,0.06525422467125787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,8192,2.2138000064425998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,512,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,128,0.055941330062018506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,1536,0.11165600352817112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,1024,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,12288,0.01977511081430647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,2560,0.04595822095870972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,256,0.0037964445849259696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,2560,0.12304977575937907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,6144,0.010236444572607676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,1536,0.02017333275742001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,16384,0.08075111442142062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,7168,0.0352151095867157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,16384,0.04309333364168803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,10240,0.07719733317693074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,51200,0.06926666365729438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,5120,0.1504862176047431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,3072,0.021354667014545862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,6144,0.025263999899228413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,1536,0.08124978012508817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,10240,0.08326133092244466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,2048,0.007076444725195567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,6144,0.09316177500618829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,512,0.060956445005204946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,2048,0.009540444446934594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,12288,0.033041778537962176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,8192,0.1312266720665826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,128,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,512,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,512,0.011242667006121742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,51200,2.6866327921549478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,3584,0.01616533266173469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,5120,0.4313208791944716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,3584,0.015757333901193406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,32,0.002867555452717675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,32,0.015953777564896476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,512,0.0037564444873068067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,2560,0.10864088932673137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,512,0.0057608890864584185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,128,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,8192,0.1592000060611301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,32,0.0028408887899584244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,65536,0.08737778001361424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,768,0.008384889198674096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,1536,0.01130488928821352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,32,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,1024,0.04226400123702156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,2048,0.01144266708029641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,1536,0.04820000131924947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,512,0.005836444596449534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,128,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,51200,0.41275821791754824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,32,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,16384,0.05923555294672648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,2560,0.00927466650803884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,5120,0.28192000918918186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,3072,0.013855111267831592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,7168,0.05379555622736613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,256,0.0032319999817344877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,2048,0.06880355543560453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,4096,0.006207110981146495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,2560,0.007711110843552484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,2048,0.2510773340861003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,2048,0.2807653215196398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,51200,0.06212977568308512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,512,0.005486222190989389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,4096,3.166910171508789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,768,0.008889777792824639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,4096,0.021260443660948012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,51200,0.018526222970750596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,10240,0.043506665362252124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,6144,0.9269484413994683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,2048,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,32,0.005117333183685939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,768,0.003378666730390655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,4096,0.027726221415731642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,4096,0.006695999867386288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,7168,0.019799111617936026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,3072,0.09752533170912002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,2560,0.16431466738382974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,32,0.024697777297761705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,12288,0.019103111492262948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,7168,0.0229111115137736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,6144,0.020610666937298246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,1024,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,16384,0.07670488622453478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,1024,0.004892444445027245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,128,0.004869333157936732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,32,0.00263288886182838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,6144,0.018405333161354065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,1024,0.08268355660968356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,10240,0.013547555440002017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,16384,1.1112115648057725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,4096,0.29175734519958496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,1536,0.29323021570841473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,7168,0.2858560085296631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,10240,0.12292533450656468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,4096,0.10946933428446452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,16384,0.023758222659428913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,12288,0.03559289044804043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,51200,0.4410684373643663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,2560,0.014462222655614218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,64,0.006033777776691649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,6144,0.011600889265537262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,12288,0.17786222034030488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,16384,0.2702239884270562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,6144,0.09272178014119466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,4096,0.05150666501786974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,16384,1.0611190795898438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,5120,0.14325422710842556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,2048,0.034296890099843345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,1536,0.014073777529928418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,768,0.005500444521506627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,3072,0.058132443163130015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,10240,0.017469333277808297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,3072,0.010199999643696679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,128,0.003206222214632564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,3584,0.09836355845133464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,1024,0.006133333262470033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,12288,0.017305778132544625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,10240,0.6118373341030544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,1536,0.012716444830099741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,256,0.00479644454187817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,1536,0.09440088934368557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,3072,0.008902221918106079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,10240,0.4323502116733127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,512,0.007899555895063613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,8192,0.021750221649805706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,10240,0.02014755540423923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,256,0.008169777691364288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,51200,0.06935822301440768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,768,0.004678222040335338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,3072,0.02358755634890662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,3584,0.07164711422390409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,51200,3.449289745754666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,16384,0.049554665883382164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,768,0.010544000400437249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,3584,0.02613777750068241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,16384,2.4815165201822915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,1536,0.010304888917340172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,768,0.07051199674606323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,8192,0.018406222263971966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,1024,0.012338666452301873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,64,0.002704888789190186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,6144,0.2735288937886556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,1536,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,128,0.0343866679403517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,12288,0.22279111544291177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,65536,0.09014311101701523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,12288,0.020415999823146395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,1024,0.006147555592987273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,10240,0.2507253222995334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,3072,0.10361599922180176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,51200,0.060927112897237144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,16384,0.5286631054348415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,1536,0.010281778044170803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,5120,0.061964445643954806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,1024,0.016781333420011733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,7168,0.18117688761817086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,6144,0.029454221328099568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,8192,0.03070755468474494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,5120,0.010018666585286459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,16384,0.20488888687557646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,5120,1.4635182486640081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,2048,0.03916622201601664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,6144,0.028195556667115953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,6144,0.014597333139843412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,12288,0.309954669740465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,12288,0.03602577911482917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,3584,0.008343111309740279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,51200,0.13848711384667292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,12288,0.24421599176194933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,128,0.002736888825893402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,51200,0.06376266479492188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,256,0.03775999943415324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,4096,0.7423271073235406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,3072,0.12117332882351345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,512,0.009363555245929295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,3072,0.03758666581577725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,65536,0.2840915520985921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,7168,0.012967999610635968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,2560,0.012500444220172035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,2048,1.037785742017958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,8192,0.02040088839001126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,7168,0.02937777837117513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,16384,0.034372445609834455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,256,0.011320888996124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,1536,0.04477599925465054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,2048,0.007131555842028723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,7168,0.04864977796872457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,65536,0.10042755471335517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,256,0.0052426668504873914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,8192,0.013973333769374423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,3072,0.046513779295815356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,64,0.0026693333768182327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,256,0.005026666654480828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,1536,0.00565422202150027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,6144,0.01019200020366245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,5120,0.17775021659003365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,32,0.0041617775956789655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,3584,0.026505778233210247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,768,0.0094355551732911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,1536,0.019853333632151287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,16384,0.02122488949033949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,65536,0.009481777747472128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,4096,0.07725955380333795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,2560,0.040251556369993426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,8192,1.5097306569417317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,512,0.004021333323584663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,3072,0.061921775341033936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,8192,0.05385955505900913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,32,0.09011110994550918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,16384,0.25584978527492946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,51200,0.07197066810395983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,1024,0.015417777829700045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,1024,0.029765334394243028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,4096,0.014594667487674289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,2048,0.006744888921578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,3072,0.026084444589085046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,1536,0.06219644678963555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,7168,4.367967817518446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,2048,0.01702311138312022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,16384,0.18774310747782388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,3072,0.026473777161704168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,6144,0.16784266630808511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,10240,0.006551110910044775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,2048,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,3584,0.10568088955349392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,1024,0.012359110845459832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,4096,0.09265066517723931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,5120,0.011757333245542316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,7168,0.01147466649611791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,16384,0.0454968876308865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,12288,0.05650222301483154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,16384,0.09820977846781413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,768,0.009458666874302758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,5120,0.0763875577184889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,65536,1.0828204684787326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,64,0.004312889029582341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,8192,0.37256267335679794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,1536,0.008383110993438298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,1536,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,6144,0.055013332102033824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,128,0.008460444708665213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,12288,0.027663111686706543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,16384,1.2541280322604709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,16384,0.860099580552843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,32,0.0028799999919202593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,2560,0.0063768890168931745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,3072,0.7487635612487793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,16384,0.3039786550733778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,512,0.005502222312821283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,1024,0.026072000463803608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,64,0.1450373331705729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,1024,0.012638222012254926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,32,0.002962666667169995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,5120,0.25819378428988987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,1536,0.004864000197913912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,768,0.010995555255148145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,12288,0.015140444040298462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,3584,0.9560924106174045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,256,0.1057271096441481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,768,0.005584888988071018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,2560,0.009769777456919352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,32,0.12064888742234971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,2560,0.16427111625671387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,2560,0.007559110720952352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,512,0.011815999945004782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,32,0.07354933023452759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,51200,2.474206288655599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,10240,0.046513779295815356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,1536,0.04713333315319485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,32,0.015572445260153877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,1536,0.04185600082079569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,512,0.0063288890653186375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,6144,0.024831111232439678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,1024,0.057836446497175426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,7168,0.015386665860811869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,6144,0.01369866645998425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,2048,0.09397244453430176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,2560,0.7200195524427625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,768,0.005836444596449534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,6144,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,5120,0.03087110983000861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,1024,0.010284444524182213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,2048,0.01144088887506061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,128,0.0046160001721647055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,1024,0.01483022173245748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,65536,1.4596320258246527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,512,0.010047111246320937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,65536,0.1251128911972046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,4096,0.009000000026490953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,51200,1.6572240193684895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,32,0.0027919999427265594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,8192,0.0664293302430047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,1024,0.04194133480389913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,51200,0.19434221585591635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,2048,0.005711110929648082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,512,0.012376888758606382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,512,0.012631110846996307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,8192,0.02459466622935401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,4096,0.18617955843607584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,7168,0.03740888833999634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,4096,0.04734933376312256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,768,0.006810666786299811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,10240,0.02274133265018463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,2048,0.006601777755551868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,512,0.025452444950739544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,4096,0.015985776980717976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,512,0.027513777216275532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,1024,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,2560,0.16230400403340658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,3584,0.09025600221421982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,65536,1.0889067120022244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,1024,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,128,0.004536889079544279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,7168,0.007674666742483775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,2560,0.01350488927629259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,2048,0.012319111161761813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,6144,0.07661688990063138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,64,0.0026622222115596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,10240,0.021287999219364587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,51200,0.060028447045220264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,16384,1.134759161207411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,768,0.06076355775197347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,4096,0.017664889494578045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,1536,0.0976453357272678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,8192,0.04275822308328417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,16384,0.7606462372673882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,1024,0.022745778163274128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,10240,0.37594132953219944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,1024,0.018312000566058688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,65536,24.586471557617188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,3072,0.00926399976015091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,32,0.013103110922707452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,7168,0.12432800398932563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,65536,84.57361518012154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,16384,0.4257342285580105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,51200,0.29455823368496364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,51200,0.08693333466847737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,51200,0.42158932156032985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,768,0.004946666873163647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,2560,0.02635555631584591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,1536,0.008230222596062554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,128,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,256,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,4096,0.011298666397730509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,10240,0.8260835541619195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,8192,0.015648888217078317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,64,0.003887999802827835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,4096,0.0129439996348487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,768,0.02343911098109351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,64,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,768,0.003555555724435382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,7168,0.012123555772834353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,5120,0.01219555570019616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,2048,0.0540906654463874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,65536,4.014364454481337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,2048,0.09936355882220799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,256,0.0035297779573334586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,65536,2.86553594801161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,51200,0.19027377499474418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,8192,0.00620266670982043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,2048,0.04680000080002678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,2048,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,32,0.04823999934726291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,6144,0.01750666730933719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,512,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,4096,0.05844710932837593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,1536,0.007486222518814935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,128,0.012789333032237159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,256,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,32,0.0030462222380770575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,2048,0.13693955209520128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,512,0.0046880000995265115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,128,0.00519466648499171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,256,0.008754666480753157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,8192,0.157423112127516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,1024,0.010646222366227044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,3072,0.12794489330715603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,2560,0.03465955456097921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,1024,0.006269333263238271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,768,0.00775999989774492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,1024,0.01000711073478063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,4096,0.012137778103351593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,128,0.03494933247566223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,4096,0.17878399954901802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,65536,1.5564586851331923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,32,0.03184533450338576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,8192,0.016723554995324876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,256,0.0031226666437255014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,16384,0.021387555532985266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,2048,0.006554666492674086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,3584,0.11184888415866429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,2560,0.601553757985433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,6144,0.16551289293501112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,3072,0.057428446080949574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,32,0.002979555477698644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,2560,0.020250666472646926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,51200,0.414264890882704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,10240,3.6603991190592446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,1536,0.28389154540167916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,3584,0.013979555832015144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,768,0.09618666436937119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,10240,0.05612533622317844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,12288,0.21217421690622965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,1024,0.00655377780397733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,8192,0.006847111301289664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,768,0.27165155940585667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,64,0.003765333443880081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,51200,0.154856006304423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,6144,0.10805155833562215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,12288,0.34297066264682347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,64,0.007040888898902469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,7168,0.025616000096003216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,1024,0.006228444476922353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,10240,0.026995556222067937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,768,0.033949332104788885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,32,0.017234666479958426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,4096,0.23386043972439238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,5120,0.1859226624170939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,3072,0.011530666715568967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,5120,0.02623733381430308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,12288,0.07695733176337348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,8192,0.015582222077581616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,4096,0.06445066796408759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,1024,0.010660444696744284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,2560,0.011777777638700275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,5120,0.21278399891323516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,7168,0.15038577715555826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,5120,0.034854223330815635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,65536,0.14949156178368464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,64,0.0026453334010309647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,128,0.025920889443821375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,51200,0.1609439982308282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,5120,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,768,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,1536,0.04304888844490051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,2048,0.029003553920321997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,1024,0.007186666958861881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,12288,0.015107555521859063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,3072,0.015966221690177917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,1536,0.03442222211096022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,32,0.13124177190992567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,256,0.004658666749795278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,1536,0.04091377721892463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,3072,0.09603910975986057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,5120,0.011740444435013665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,512,0.015808888607554965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,1024,0.006138666636413998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,1536,0.07739466428756714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,256,0.006989333364698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,12288,0.05393155415852865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,12288,0.06905955738491483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,16384,17.969136555989582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,256,0.004541333350870344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,32,0.01889866590499878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,5120,0.10659111208385891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,7168,0.09975200229220921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,4096,0.12123200628492568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,2560,0.05076622300677829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,64,0.007174222005738153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,256,0.0672524438963996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,256,0.0031066667288541794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,8192,5.174634721544054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,4096,0.014488000008794995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,7168,0.028943998946083918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,8192,0.12680178218417698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,7168,0.007072889142566257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,4096,0.019333332777023315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,12288,0.019499555230140686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,1024,0.006134222365087933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,6144,0.0938897795147366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,10240,0.030596445004145306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,64,0.005205333232879639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,768,0.13305421670277914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,32,0.003657777690225177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,51200,18.698462592230904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,3584,0.011446221835083432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,3584,0.022805333137512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,64,0.008974221845467886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,12288,0.16115821732415095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,8192,0.015311110350820752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,3584,0.010867555936177572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,16384,0.11098400089475845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,12288,0.36120978991190594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,1536,0.007308444215191736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,5120,0.05252444412973192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,7168,0.03200711144341363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,7168,0.3143591086069743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,16384,0.023747555083698694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,65536,0.11988266309102376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,10240,0.013812444276279874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,2560,0.01752355529202355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,4096,0.06667555703057183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,32,0.0026559999419583213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,65536,0.20293154981401232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,3072,0.00926577796538671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,5120,0.016661332713233102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,3584,0.010424888796276517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,51200,0.10858933130900066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,3584,0.09260355763965183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,768,0.030076444149017334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,1024,0.02574311031235589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,12288,0.03331022130118476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,10240,0.15399377875857884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,5120,0.014769777655601501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,256,0.028636445601781208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,256,0.007912000020345053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,51200,0.03592355714903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,512,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,7168,0.07301688856548734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,10240,0.016117332710160148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,4096,0.10849066575368245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,1536,0.007241778075695038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,16384,0.02474044428931342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,10240,0.01883288886811998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,10240,0.5247751341925727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,2560,0.008150222400824228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,32,0.0032479998966058097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,768,0.00370133337047365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,32,0.004910222358173794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,4096,0.014327110515700446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,128,0.0036311112344264984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,512,0.04080355498525832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,768,0.022085333863894146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,6144,0.028279112444983587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,2048,0.0057608890864584185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,2048,0.006199999815887875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,5120,0.30764712227715385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,768,0.23688888549804688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,768,0.007485333416197035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,128,0.00507377791735861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,7168,0.8791759808858236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,10240,0.04092266824510362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,5120,0.26324354277716744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,3584,0.05485511157247755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,6144,0.200673778851827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,3072,0.06025333536995781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,1536,0.09412178066041733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,2560,0.03742400142881606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,65536,0.19801955752902559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,256,0.03447555502255758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,10240,0.01368977791733212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,1024,0.006162666612201267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,256,0.06056444512473213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,768,0.006271111054552927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,3584,0.04974488748444451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,4096,0.07530755466885038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,65536,0.09815288914574517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,51200,5.4293670654296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,2560,0.0499582224422031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,768,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,128,0.00812977800766627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,1536,0.029994666576385498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,5120,0.044960889551374644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,51200,0.06408711274464925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,512,0.0033155555526415506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,512,0.005555555638339784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,2560,0.018615111708641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,2048,0.012362666428089142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,3584,0.008750222623348236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,2048,0.01404355631934272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,3072,0.06795289119084676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,8192,0.1160702175564236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,12288,0.09523644712236191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,16384,0.047840889957216054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,256,0.037435554795795016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,1024,0.22943200005425346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,3072,0.02212000058756934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,65536,0.07041422526041667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,7168,0.39249245325724286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,768,0.00902133352226681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,1024,0.012070222033394707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,16384,0.024870221813519795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,3072,0.041803555356131665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,3072,0.0965848896238539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,51200,0.1741244395573934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,256,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,51200,3.8155636257595487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,6144,0.20332888762156168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,128,0.009898666706350114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,8192,0.06701333655251397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,1536,0.005786666853560342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,1536,0.3908409012688531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,51200,2.887700398763021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,128,0.004132444245947732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,65536,0.17796711126963297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,3072,0.02443644404411316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,1536,0.019876443677478366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,3584,0.007958222594526079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,256,0.017102221647898357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,1536,0.044807109567854136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,256,0.010018666585286459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,512,0.5459946526421441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,128,0.009347555538018545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,16384,0.13508089383443198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,7168,0.008163555628723567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,8192,0.017666666044129264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,2048,0.012665777570671506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,256,0.012288889123333825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,768,0.005853333406978183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,128,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,3584,0.008952000074916417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,6144,0.011964444484975604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,3584,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,51200,0.12721067004733616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,1024,0.013782222237851886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,8192,0.49807198842366535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,1024,0.0741457806693183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,2560,0.04002755549218919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,64,0.004927111168702443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,2560,0.050865779320398964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,2560,0.008167111211352879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,5120,0.008839111361238692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,1536,0.0063546668324205614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,6144,0.05371555685997009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,256,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,1024,0.006647111227114995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,65536,0.12141244941287571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,128,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,12288,0.4993831316630046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,32,0.004976888911591636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,128,0.006089777582221561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,1536,0.004662222332424588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,128,0.06673155228296916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,2048,0.013929777675204806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,7168,0.01333688861793942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,3584,0.008638222184446123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,7168,0.14608889155917698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,8192,0.012682666381200155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,1536,0.048956443866093956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,256,0.03561777869860331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,51200,0.009554666777451834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,51200,0.07668533590104845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,3072,0.09221688906351726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,4096,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,1024,0.005933333188295364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,8192,0.006685333533419504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,128,0.009700444837411245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,3072,0.0074924445814556545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,128,0.013076444466908773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,3072,0.015436444017622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,512,0.008775110873911116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,3584,0.02087466584311591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,2048,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,128,0.009678222239017487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,128,0.009859555297427708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,768,0.004963555683692296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,768,0.06608533196979098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,768,0.10255289077758789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,16384,0.08962488836712307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,3584,0.008190222084522247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,3072,0.021576000584496394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,512,0.009842666486899057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,512,0.040716442796919085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,1024,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,7168,0.256494230694241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,2048,0.011819555527634092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,10240,0.605113771226671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,6144,0.011739555332395764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,12288,0.03553244471549988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,64,0.003489777859714296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,3072,0.10976711246702407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,32,0.003551111039188173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,1024,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,256,0.0033493333806594214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,1024,0.005176888985766305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,3584,0.18967200650109184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,12288,4.049872080485026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,3584,0.012039110892348819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,10240,0.10894399881362915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,512,0.004041777716742622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,768,0.004037333445416557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,768,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,2560,0.020627554919984605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,12288,0.2606755627526177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,64,0.006266666783226862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,1024,0.004265777766704559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,3584,0.04212177793184916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,12288,1.0870595508151584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,6144,0.22404800521002877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,768,0.014256000518798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,2560,0.014101333088344999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,5120,0.2439191076490614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,64,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,512,0.019347555107540555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,128,0.006076444354322221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,16384,0.011672000090281168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,2560,0.008575110799736446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,65536,0.12350577778286403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,10240,0.16711644331614176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,65536,0.08445244365268284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,51200,0.7131982379489474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,6144,0.048047112094031445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,8192,0.0318195554945204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,64,0.020768889122539096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,3072,0.06898044215308295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,128,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,3072,0.04299644298023648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,51200,0.21748711003197563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,512,0.02442933287885454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,5120,0.027079999446868896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,256,0.010249777800507015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,3072,0.015095111396577625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,1536,0.009125333693292405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,2560,0.022983110613293115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,10240,0.09272888633939956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,512,0.01687022215790219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,65536,0.07367733452055189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,8192,0.01687733332316081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,2048,0.025099555651346844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,4096,0.012541333006487953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,5120,0.01680799987581041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,7168,0.01498311095767551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,32,0.0029120000286234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,12288,0.022310222188631695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,256,0.010454222559928894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,51200,0.11082133319642808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,1536,0.00847466703918245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,3072,0.008818666968080733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,10240,0.12685688336690268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,51200,0.20960711108313668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,1536,0.017151110702090792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,2048,0.009039999710188972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,2048,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,128,0.00300177786913183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,2560,0.19942311445871988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,5120,0.09363200267155965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,12288,0.017765333255132038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,1024,0.058351112736595996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,256,0.0046773333516385825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,10240,0.00942133367061615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,768,0.03970400161213345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,5120,0.1408568885591295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,4096,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,4096,0.006248000181383557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,8192,0.17528533935546875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,16384,0.10340266757541233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,12288,0.17154222064548066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,2560,0.00796088907453749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,5120,0.017294221454196505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,3072,0.018764444523387484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,6144,0.01945688823858897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,3072,0.1958062251408895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,64,0.0024195555597543716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,768,0.005937777873542573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,5120,0.08204622401131524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,768,0.056537780496809215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,12288,0.04222311245070564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,12288,4.942622290717231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,1536,0.006169777777459886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,8192,0.6218844519721137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,1536,0.0075804442167282104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,256,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,768,0.005693333430422678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,12288,0.43684710396660703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,51200,0.07915644513236152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,8192,0.1413688924577501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,1536,0.006513777706358168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,2048,0.007827555967701806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,512,0.04126311010784573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,3584,0.03472000029351976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,16384,0.009039999710188972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,4096,0.031685332457224526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,2048,0.02896711230278015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,51200,0.06996355454126994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,256,0.0047386665311124595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,256,0.0052577778697013855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,3584,0.09783022271262275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,256,0.1187395519680447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,64,0.006787555499209299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,10240,0.02572888963752323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,2560,0.010932444698280759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,768,0.025250666671329077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,1024,0.044759111271964185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,1024,0.025921778546439275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,8192,0.011569778124491373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,51200,0.2827013333638509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,3584,0.018581334087583754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,3584,0.01013777818944719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,4096,0.03670933180385166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,5120,0.22581511073642305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,2048,7.673505995008681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,512,0.026039999392297532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,10240,0.4531715710957845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,51200,1.1300239562988281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,32,0.008946666287051307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,1024,0.011394666300879585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,3584,0.01960444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,3584,0.03840355409516229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,16384,0.22917776637607148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,3072,0.027477333943049114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,5120,0.03688622183269925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,10240,0.006484444356626934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,51200,0.11117511325412327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,1536,0.03874399926927354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,65536,0.3430666658613417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,2048,0.022276444567574397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,4096,0.019928889142142404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,512,0.008768888811270395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,768,0.004295999805132548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,12288,0.049680888652801514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,6144,0.09442933400472005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,768,0.007141333487298753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,1024,0.005021333280536863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,1536,0.042378667328092784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,51200,0.13196533256106907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,1024,0.034838222795062594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,64,0.010729778144094678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,51200,0.22166221671634248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,10240,0.09851111306084527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,5120,0.013314666847387949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,5120,0.030576888057920668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,512,0.005408888889683618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,768,0.00906399968597624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,65536,0.08371288908852471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,6144,0.21083554956648085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,3072,0.015115555789735583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,12288,0.25992711385091144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,128,0.008434666527642144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,65536,0.13157067033979628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,2560,0.011928000383906893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,5120,0.11036444372600979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,3072,0.06772622134950426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,64,0.005212444398138259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,7168,0.07320177555084229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,10240,1.2056692971123588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,3584,0.4285413424173991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,5120,1.794040044148763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,3072,0.00849599970711602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,1536,0.04851822058359782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,3072,0.010999999940395355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,3584,0.04108355442682902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,2048,0.035248001416524254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,4096,0.08676266670227051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,3584,0.018986667195955913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,2048,0.15123289161258274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,51200,3.034194734361437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,3584,0.017297777864668105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,3584,0.012562666502263812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,3072,0.0527688894006941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,6144,0.025859556264347498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,1024,0.039671109782324895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,51200,0.05785333447986179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,16384,0.06526577472686768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,5120,0.013930666777822705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,8192,0.012445333103338877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,16384,0.19029155042436388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,1024,0.005668444352017508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,51200,0.05558933152092827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,10240,0.12967644797431097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,768,0.023036445180575054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,3072,0.06688533226648967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,4096,0.23506400320265028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,1024,0.007090667055712805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,65536,0.014064000712500678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,8192,0.014567111929257711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,5120,0.05152533451716105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,51200,0.06975466675228544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,7168,0.010433778166770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,2048,0.006890666567616993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,10240,0.04894933435651991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,3584,0.01126844435930252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,3072,0.009518221848540837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,2048,0.0064195555945237475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,1024,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,1024,0.0040942223535643685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,8192,0.21139910486009386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,128,0.007093333535724216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,12288,0.015989333391189575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,64,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,4096,0.00888888869020674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,512,0.04381422201792399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,3072,0.16720978418986002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,32,0.0047128887640105355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,3584,0.016898666818936665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,10240,0.16584711604648167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,2048,0.0343466666009691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,1024,0.0036915557252036203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,16384,0.06842933098475139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,4096,0.1632711092631022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,16384,0.006171555568774541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,64,0.008825778133339351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,1536,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,3584,0.0073928890956772705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,2048,0.011047999891969891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,2048,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,32,0.005039999882380168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,3072,0.015967999895413715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,16384,0.026872000760502283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,512,0.006760889043410619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,1536,0.010671999719407825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,256,0.014848000473446317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,256,0.057586669921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,1024,0.010302222437328762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,1024,0.010022222167915767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,768,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,12288,0.21535555521647134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,768,0.020594666401545208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,64,0.0026559999419583213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,2048,0.014785778191354541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,2560,0.04015288750330607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,4096,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,32,0.0052284445199701525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,4096,0.021731555461883545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,1024,0.02131822208563487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,10240,0.07626399728986952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,512,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,8192,0.008749333520730337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,7168,0.10394044717152913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,3584,0.018089777893490262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,6144,0.02499733368555705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,2560,0.09847111172146267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,1024,0.017456000049908955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,128,0.005074666606055366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,51200,25.013307359483505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,512,0.011599111060301462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,128,0.026439110438028973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,8192,0.013197333448463015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,2560,0.04933333396911621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,1024,0.007071110937330458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,2048,0.009098666409651438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,512,0.0262080000506507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,256,0.02954933378431532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,7168,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,64,0.0025191110455327565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,5120,0.11818844742245144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,51200,0.17414755291408965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,1024,0.011696889168686338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,3072,0.029554665088653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,4096,0.010416888528399997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,10240,0.08580266767077976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,512,0.003579555700222651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,1024,0.007223111059930589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,51200,0.2708631091647678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,512,0.005484444399674733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,768,0.040123555395338274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,7168,0.26330577002631295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,5120,0.013061333033773633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,1024,0.01351377781894472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,3072,0.04554933309555054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,6144,0.02801333202256097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,51200,1.5893440246582031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,3072,0.33540977372063535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,64,0.005417777846256892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,512,0.18025067117479113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,12288,0.01961866683430142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,3584,0.010108444425794814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,1024,0.011736888852384357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,1024,0.06459644767973158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,1024,0.009516444471147325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,2048,0.01706133286158244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,768,0.03485777642991807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,16384,0.20646488666534424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,8192,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,2560,0.0787715580728319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,32,0.005032888717121548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,128,0.02425244450569153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,3584,0.009504888620641496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,6144,0.0337084432442983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,64,0.011872889267073737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,1536,0.006789333290523953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,2048,0.01147466649611791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,16384,0.031261334816614784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,512,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,10240,0.0236435549126731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,12288,0.024159111910396155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,2560,0.038803554243511625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,32,0.004469333423508538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,51200,0.008126222425036961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,128,0.005052444421582752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,1536,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,2560,0.012443555725945367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,5120,1.1695973078409831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,6144,0.20041422049204508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,5120,0.011613333390818702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,7168,0.2026862170961168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,10240,1.6687759823269313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,512,0.006849777781301075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,1536,0.01016177816523446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,128,0.004615111069546806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,64,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,5120,0.016334222422705758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,32,0.017250667015711468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,512,0.016688000824716356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,256,0.006009777800904379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,65536,0.12422666284773086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,6144,0.0097884444726838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,8192,0.052501334084404834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,1536,0.006767999794748094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,1024,0.00998133338159985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,51200,3.950778537326389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,12288,0.022932445009549458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,2048,0.009282666775915358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,3072,0.5107759899563259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,10240,0.04172799984614054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,1536,0.007857778006129796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,7168,0.1550604502360026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,2048,0.00758844448460473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,1536,0.8435884581671821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,4096,0.041343109475241765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,1024,0.023359111613697473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,1536,0.4050977759891086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,1024,0.0053839998112784485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,1536,1.2286870744493272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,10240,0.1876337793138292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,2560,0.00831911133395301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,51200,0.4649048911200629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,32,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,32,0.006940444310506185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,8192,0.012257777982287936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,1536,0.07379733191596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,8192,0.35202754868401426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,7168,0.014939554863505893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,16384,4.922751108805339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,8192,0.2360995610555013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,2560,0.013320888910028668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,12288,0.17656355433993867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,16384,6.6812184651692705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,3072,0.009124444590674506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,32,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,8192,0.015683554940753512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,8192,0.0258586671617296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,3584,0.28106845749749076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,1536,0.03565333286921183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,7168,0.7922204335530599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,768,0.012789333032237159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,3584,0.11714933978186713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,4096,0.03580977850490146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,16384,0.07763910955852933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,16384,0.11398222711351182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,256,0.004876444323195351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,3072,0.007535999847782983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,10240,0.016137777103318107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,2560,0.014544889330863953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,4096,0.058893336190117725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,7168,0.04105066590838962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,128,0.011133333047231039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,7168,0.2673111226823595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,512,0.007673777639865875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,3584,0.08627377616034614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,2048,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,768,0.0036728887094391715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,10240,0.0425582230091095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,51200,0.05788889196183947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,7168,0.5483448770311143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,8192,0.5405111312866211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,2560,0.00887288898229599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,3584,0.020216888851589628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,768,0.00645955569214291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,51200,0.29960621727837455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,5120,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,2048,0.6047680113050673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,128,0.019325332509146798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,2560,0.08699911170535618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,6144,0.0162462227874332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,4096,0.01625866691271464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,5120,0.00830666638082928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,1536,0.0064071110553211635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,128,0.011499555574523078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,768,0.00629066675901413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,768,0.11613778273264568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,128,0.008654221892356873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,3584,1.1201946470472548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,7168,0.02535111043188307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,256,0.0038151111867692736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,7168,0.02702222267786662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,5120,0.014843554960356818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,512,0.0063768890168931745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,4096,0.014858666393491956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,2048,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,768,0.007899555895063613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,8192,1.1745902167426217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,6144,0.046462221278084644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,4096,0.011730666789743634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,3072,0.039763556586371526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,5120,0.015172445111804538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,64,0.0034755555291970572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,1024,0.008105778031879002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,3072,0.01034577770365609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,8192,0.08726933267381455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,65536,0.07220977544784546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,32,0.007997333175606197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,10240,0.018889778190188937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,5120,0.03494133220778571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,64,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,12288,0.016711110870043438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,65536,0.25436533821953666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,1024,0.1256764464908176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,5120,0.06096533272001478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,768,0.0898053314950731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,32,0.0029484445436133277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,16384,0.13454400168524847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,51200,0.19713955455356172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,128,0.006761777732107375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,51200,13.906169467502169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,7168,0.04830133252673679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,12288,0.1378693315717909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,65536,0.2970080110761854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,2048,1.2034515804714627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,512,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,4096,0.13241777155134413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,7168,0.18496533234914145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,51200,2.4089014265272355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,2560,0.012819555898507437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,2048,0.0669706662495931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,51200,1.9696231418185766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,768,0.011208888557222156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,6144,0.1256248950958252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,16384,0.006325333482689328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,4096,0.15597600407070583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,512,0.005391110976537068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,3072,0.1756604380077786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,16384,0.03569511241383023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,4096,0.8256862428453234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,3584,0.009201777478059133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,12288,0.052264889081319175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,16384,0.02657777733272976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,65536,0.406444443596734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,5120,0.01532444523440467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,10240,0.018776888648668926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,51200,0.7174417707655164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,64,0.04368444283803304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,128,0.002924444360865487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,256,0.030724445978800457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,10240,0.11476622687445746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,7168,0.026797332697444495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,3072,0.014064888159434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,10240,0.02902488907178243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,51200,0.06991289059321086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,256,0.014691554837756686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,51200,10.405371771918402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,256,0.043106665213902794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,1024,0.020115555988417733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,6144,0.7049742274814181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,7168,0.3181893295711941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,4096,0.011976888610257043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,12288,0.5419342252943251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,1536,0.25675466325547963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,65536,4.028071933322482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,256,0.007713778151406183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,256,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,128,0.05315999852286445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,768,0.015271110667122735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,6144,0.011184000306659274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,5120,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,1536,0.01051111105415556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,64,0.006205333189831839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,7168,0.3274853229522705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,3584,0.018239999810854595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,65536,0.8889670901828342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,256,0.006492444624503453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,2560,0.07460444503360324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,8192,0.011732444167137146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,8192,0.01715377800994449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,4096,0.02790044413672553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,2560,0.009940444595283931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,128,0.005095110999213325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,64,0.0026293332791990707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,1536,0.0069226668112807805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,65536,1.0440702438354492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,1536,0.007455999652544658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,3072,0.23061688741048178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,128,0.003491555651028951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,16384,0.014529777897728814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,256,0.03859733210669623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,768,0.01200977795653873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,64,0.008364444805516137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,6144,0.021930666433440313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,16384,0.10626311434639825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,65536,2.705337736341688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,512,0.1675342188941108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,64,0.02896889050801595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,3584,0.01261333293384976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,256,0.008759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,3072,0.016366221838527255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,5120,0.013035555680592855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,51200,0.1787253353330824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,4096,0.05230933427810669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,65536,0.21862222088707817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,3584,0.04450400008095635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,32,0.0027057776848475137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,12288,0.04927466644181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,51200,0.00833777752187517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,16384,0.018533334136009216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,6144,0.05183555682500204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,16384,3.341702143351237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,768,0.010120000276300643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,5120,0.6656488842434353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,5120,0.06272444460127088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,256,0.003237333355678452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,12288,5.921686384412978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,2048,0.5423395368787978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,51200,0.7270897759331597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,51200,0.22179199589623344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,65536,0.0681155522664388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,768,0.0058542220956749385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,128,0.006933333145247565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,16384,0.009863999982674917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,64,0.005515555540720622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,3072,0.011890666352377998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,512,0.007592889169851939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,51200,1.2127866744995117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,256,0.07288266552819146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,1024,0.008825778133339351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,10240,0.1806186702516344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,6144,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,7168,0.020942222740915086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,256,0.0047191112405723995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,16384,0.06228533056047228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,10240,0.12521066930558947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,12288,0.01870933340655433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,3584,0.04680088824696011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,128,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,16384,0.046181334389580615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,7168,0.01167377746767468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,128,0.0027377777215507296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,128,0.009903999666372934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,256,0.0053324442770746015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,16384,0.01724088854259915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,2048,0.12179822391933864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,128,0.06610221995247735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,6144,0.0064693333374129395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,1536,0.006446222050322427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,2560,0.006227555374304454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,768,0.03600355651643541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,5120,0.05121244324578179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,3072,0.019169777631759644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,3072,0.008782222039169734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,8192,0.02014400064945221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,5120,0.28481777509053546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,2560,0.012096000214417776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,4096,2.802175097995334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,1024,0.005987555616431766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,16384,0.013771555489963956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,12288,1.153779559665256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,3584,0.011374221907721626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,5120,0.016723554995324876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,8192,0.5082657602098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,8192,0.04435733291837904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,51200,0.1325039996041192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,1536,0.035623110002941556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,51200,0.008192888564533657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,1536,0.0052328887912962176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,512,0.003891555385457145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,1024,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,51200,1.3332381778293185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,2048,0.03839822279082404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,512,0.006720888945791457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,1536,0.004459555364317364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,256,0.004643555730581284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,1536,0.16217954953511557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,10240,0.26447823312547475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,768,0.009355555805895064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,1536,0.044144001271989614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,768,0.005587555468082428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,51200,0.22288266817728677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,2048,0.013185777597957186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,51200,1.2110666698879666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,32,0.008285333712895712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,3072,1.2333173751831055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,3072,0.02472177810139126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,51200,0.061576002173953585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,2560,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,2560,0.11454399426778157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,128,0.0028737777223189673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,4096,0.010623111493057676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,1536,0.07162755727767944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,4096,0.008648888932334052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,512,0.005393777870469623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,256,0.008671110702885522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,7168,0.027775999572541978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,4096,0.2329253355662028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,2560,0.00923911150958803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,5120,0.0372648901409573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,12288,0.07904799779256184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,1024,0.01016622202263938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,8192,0.01901955571439531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,7168,0.06735022200478448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,65536,0.38826844427320695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,51200,38.68882581922743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,3072,0.054083555936813354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,256,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,512,0.0034933334423436057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,10240,0.05128622386190626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,64,0.005547555370463266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,64,0.008658666577604082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,3584,0.2669120099809435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,12288,0.016900445024172466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,512,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,12288,0.42926843961079914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,8192,0.16004799471961126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,16384,0.16174577342139349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,32,0.011281777587201862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,8192,0.038078222009870745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,16384,0.02144977781507704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,6144,0.5083422131008571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,2048,0.04249688982963562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,10240,0.021293333835071985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,5120,0.05392533209588793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,51200,0.3772515455881755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,4096,0.009875555833180746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,2560,0.014313777287801107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,6144,0.40734222200181747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,51200,7.33438703748915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,1024,0.04211022125350105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,3584,0.04904355605443319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,1024,0.005506666584147348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,768,0.007693332930405934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,6144,0.02511111067401038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,512,0.0035111109415690103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,2560,0.02293599976433648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,3584,0.20374933878580728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,256,0.003022222262289789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,1536,0.012874667015340594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,65536,0.20150310463375518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,65536,0.20377688937717012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,65536,0.11297689543830024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,256,0.006817777951558431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,4096,0.12234666612413193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,7168,0.1083751122156779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,5120,0.01795733306143019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,3072,0.33129334449768066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,2048,0.01738933391041226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,1536,0.015504888362354703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,128,0.006597333484225803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,6144,0.012146666646003723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,5120,0.16188000308142766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,2048,0.006154666758245892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,512,0.003731555408901639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,32,0.0058355554938316345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,1024,0.005877333382765452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,512,0.00553955551650789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,768,0.014865777558750577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,4096,0.04808355702294243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,1536,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,64,0.09339466359880234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,32,0.003670222229427762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,65536,0.3044346703423394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,1024,0.00996977753109402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,256,0.003428444473279847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,1024,0.007082666787836287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,2560,0.009847111172146266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,10240,0.013122667041089801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,7168,0.008919110728634728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,6144,0.026314665873845417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,1536,0.007420444654093848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,128,0.0027804445061418745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,4096,0.06476089027192858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,3072,0.15424711174435085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,768,0.006058666855096817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,4096,0.0112631113992797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,2560,0.007070221834712558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,10240,0.11641244093577068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,256,0.005566222386227713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,1536,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,2560,0.009566222627957663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,3072,0.040279110272725425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,1536,0.006548444430033366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,8192,0.01165066659450531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,8192,0.10029155678219265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,64,0.008207999997668797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,5120,0.01216533366176817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,2560,0.10372889041900635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,10240,0.014759999182489185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,65536,1.3049369388156467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,5120,0.038350221183564924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,1536,0.0354728897412618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,256,0.005676444619894028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,6144,0.012961777547995249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,3584,0.01567911108334859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,7168,0.13410845067765978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,512,0.04931733343336317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,51200,0.31337955262925893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,10240,0.021370665894614324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,1536,0.03678311241997613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,12288,0.3514373302459717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,3072,0.031182222896152075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,7168,0.09063200155893962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,65536,1.1337671279907227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,512,0.005596444424655702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,10240,0.39770489268832737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,7168,0.02968977888425191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,7168,0.09226133426030476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,1024,0.03635911146799723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,10240,0.16678222020467123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,6144,0.011072888970375061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,32,0.007100444700982835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,7168,0.01443822185198466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,768,0.0063279999627007385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,5120,0.01072622256146537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,512,0.01457155578666263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,4096,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,10240,2.370567957560221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,3584,0.02587022218439314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,128,0.00980977796845966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,4096,0.055664890342288546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,4096,0.008464888566070134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,1536,0.005517333332035277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,51200,0.06939466794331868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,1536,0.06392977635065715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,16384,0.06214400132497152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,5120,0.013244444297419654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,5120,0.01925511161486308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,128,0.009107555780145857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,10240,0.019034667147530448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,256,0.006375999914275275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,3584,0.06818399826685588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,3072,0.02260977857642704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,2560,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,128,0.01534844438234965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,5120,0.022493332624435425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,5120,0.022706665926509436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,4096,0.010149333212110732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,7168,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,8192,0.17588444550832114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,32,0.003039999968475766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,512,0.0047022220161226065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,7168,0.016947555873129103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,12288,1.925098631117079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,32,0.0033973333322339584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,3072,0.07560711436801486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,768,0.006180444525347815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,5120,0.21490132808685303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,6144,2.0440648396809897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,64,0.012015110916561551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,256,0.01550222271018558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,768,0.00342399999499321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,2048,0.01458666721979777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,2048,0.6437342431810167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,5120,0.01611288885275523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,7168,0.14256978034973145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,3584,0.02085777786042955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,1024,0.003969777789380816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,4096,0.01571555601225959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,2560,0.013979555832015144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,7168,0.020946666598320007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,10240,0.03936266567971971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,65536,0.07083200083838569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,5120,0.009011555876996782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,3072,0.014952888091405233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,2048,0.774895985921224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,32,0.003105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,4096,0.05221244361665514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,6144,0.11092533005608453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,51200,3.8036354912651906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,5120,0.012455110748608908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,3584,0.09622666570875381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,1024,0.059044447210099965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,7168,0.3272533416748047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,3584,0.008632889224423302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,2048,0.01834222177664439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,4096,0.08212800158394708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,128,0.00328533330725299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,256,0.007191999918884701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,512,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,128,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,10240,0.9312338299221463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,128,0.004456888884305954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,1536,0.007658667034573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,1536,0.006047111004590988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,65536,8.70548333062066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,4096,0.10310222042931451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,3584,0.08582488695780437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,2048,0.9417671627468533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,51200,58.38573540581597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,65536,0.10078222221798366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,512,0.011722666521867117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,3072,0.0420595539940728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,65536,0.07605066564348009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,7168,0.0395760006374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,3072,0.017338666650984023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,768,0.036511109934912786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,8192,0.09361066420873006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,10240,0.2410435676574707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,1024,0.006116444451941385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,128,0.004209777961174647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,16384,0.08202755451202393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,1536,0.012695111334323883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,5120,0.014993778533405729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,3584,0.05531555414199829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,4096,0.0088755554623074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,2048,0.011454222102959951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,3072,0.008775110873911116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,256,0.011778666741318174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,1024,0.03958755400445726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,51200,0.1708702246348063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,32,0.002806222273243798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,256,0.00369422220521503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,512,0.018664888209766813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,4096,0.16068977779812282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,16384,3.3076089223225913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,256,0.005686222265164058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,1024,0.011396444506115384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,768,0.004319111092223061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,2048,0.029637333419587877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,768,0.10904533333248562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,6144,2.005023956298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,1536,0.010267555713653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,256,0.007110222346252865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,1536,0.024900444679790076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,1024,0.004307555655638377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,64,0.0047866664826869965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,1024,0.004043555508057277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,4096,0.018005333013004728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,512,0.01938577824168735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,10240,0.013622221847375235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,1536,0.012846222354306115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,2560,0.03165422214402093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,1024,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,3584,0.008966222405433655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,5120,0.10156710942586263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,3584,0.010503999888896942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,32,0.005125333451562458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,8192,0.5267199940151638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,12288,0.006575999988449945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,1024,0.1117280059390598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,65536,0.18617867098914254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,1536,0.006687110910813014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,2560,0.1446311076482137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,1024,0.01349422252840466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,3072,0.022488888767030504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,6144,0.014262222581439547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,1536,0.03998044464323256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,16384,1.049247105916341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,512,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,32,0.0029253332565228143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,512,0.00618577789929178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,3584,0.12751644187503391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,51200,0.08533066511154175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,10240,0.010593777729405297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,12288,0.03008355696996053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,32,0.004107555581463708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,768,0.08823466963238186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,5120,0.02496533261405097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,10240,0.2725191116333008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,4096,0.22099378373887804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,12288,6.743290795220269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,1024,0.007689778175618913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,10240,0.06987555821736653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,10240,0.010595555934641095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,8192,0.01979200045267741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,1024,0.012874667015340594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,2048,0.008671999805503422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,6144,0.0251866661840015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,65536,0.13468533092074925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,10240,0.27760712305704754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,2048,0.015261333849694995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,3072,0.009027555584907532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,2560,0.020071110791630216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,512,0.01347733371787601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,3584,0.008400888906584846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,10240,0.030344890223609075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,3072,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,1536,0.01278133359220293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,1536,0.008183110919263627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,512,0.005769777629110549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,6144,0.12284799416859944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,8192,0.015622221761279635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,6144,0.041872888803482056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,51200,0.04483644498719109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,51200,0.1409777800242106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,8192,0.157185779677497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,65536,0.3726657761467828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,3584,0.013148444394270578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,2560,0.11178133222791885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,768,0.0074026667409473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,51200,0.08185688654581706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,512,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,2048,0.0184542222155465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,7168,0.2518133322397868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,10240,0.9102587170071073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,3072,0.09358666340510051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,256,0.006549333532651265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,5120,0.3551573223537869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,12288,0.01963200006220076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,1536,0.0984862248102824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,65536,0.07932621902889676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,3072,0.0594959987534417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,5120,0.0184568895234002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,3072,0.03195910983615451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,3584,0.00904444439543618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,5120,1.2335529327392578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,768,0.007295110987292395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,768,0.006024000131421619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,7168,0.04153688748677572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,10240,0.018329777651362948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,2560,0.01477777792347802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,3072,0.02179644505182902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,64,0.027621333797772724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,51200,0.015119110544522604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,2560,0.007234666910436418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,16384,0.011542222566074796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,3072,0.02458488941192627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,7168,0.12159555488162571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,10240,0.03735555542839898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,65536,0.1337946653366089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,2560,0.006686222222116258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,768,0.0038746665749284956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,1536,0.005620444400442972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,4096,0.02442488902144962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,1024,0.009382222261693742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,1536,0.08257777823342217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,3584,0.008239111138714684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,256,0.002806222273243798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,10240,0.023640000157886084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,256,0.01144977741771274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,2048,0.034472889370388456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,51200,0.11167377895779079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,6144,0.012224888636006249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,10240,1.7437049018012152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,2048,0.010260444548394945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,2560,0.05904710955089993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,128,0.02420799930890401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,65536,0.07767911089791192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,5120,0.022608000371191237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,8192,0.24141245418124727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,256,0.0033528889632887314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,32,0.008488000267081791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,2048,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,3584,0.027334221535258826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,3072,0.0188426673412323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,16384,0.04853777752982246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,12288,2.5399093627929688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,1024,0.014518222875065274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,2048,0.078548444641961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,2048,0.007919111185603673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,16384,0.08047822448942396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,7168,0.030504888958401147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,7168,0.035754667388068304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,6144,0.012113778127564324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,12288,0.016131555040677387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,6144,0.021856889128684998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,2048,0.006477333191368315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,6144,0.014840889308187695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,512,0.00527022240890397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,12288,4.579125298394097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,8192,0.1286462280485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,3072,0.011692444483439127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,256,0.02388622197839949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,65536,0.8424631224738227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,16384,0.4104337692260742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,16384,0.2616746690538195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,768,0.018221333622932434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,3072,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,768,0.028216001060273912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,3072,0.19360978073543975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,4096,0.06226133637958103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,6144,0.019153777096006606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,1024,0.006137777947717243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,51200,0.28486045201619464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,16384,0.047913776503668896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,32,1.0371315214369032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,51200,0.19507377677493623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,10240,0.020080889264742535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,128,0.008744888835483128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,12288,0.017457778255144756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,512,0.035290665096706815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,512,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,2048,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,8192,0.6466373337639703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,2048,0.010647999743620554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,7168,0.029896001021067303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,4096,0.25312532318962944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,5120,0.023431110713216994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,512,0.003269333392381668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,4096,0.006436444405052397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,6144,0.2897164556715223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,51200,2.1565377977159286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,128,0.010077333284748925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,51200,0.2267297903696696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,10240,0.06740266746944852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,2560,0.007476444045702617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,3584,0.0592213339275784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,10240,0.011713777979214987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,512,0.0032960000551409195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,256,0.07256622446907891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,128,0.009008888569143083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,6144,0.4064364433288574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,16384,0.17514310942755806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,32,0.008236444658703273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,3584,0.009139555195967356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,768,0.013046222428480784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,2560,0.007051555646790399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,51200,0.6361164516872829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,1024,0.025302223033375208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,1024,0.004019555532270008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,1536,0.007408888803588019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,4096,0.0919297801123725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,3584,0.3974382347530789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,51200,2.1719173855251737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,65536,0.024872000018755596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,1536,0.4085235595703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,16384,0.24622665511237252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,51200,0.014115555418862237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,7168,0.023956444528367784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,65536,0.06348977486292522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,4096,1.2495724360148113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,32,0.01880088945229848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,3072,0.02542488939232296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,4096,0.008361777497662438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,3072,0.009169778062237633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,2048,0.01186044431395001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,10240,0.06737689177195232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,2048,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,10240,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,7168,2.3267093234592013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,512,0.00397511116332478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,16384,0.04804799954096476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,3072,0.010681777364677854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,128,0.0032799999333090256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,8192,0.14309955967797175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,16384,0.03746933407253689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,768,0.005518222020732031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,12288,0.05295555459128486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,12288,0.029120888974931505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,8192,0.010026666853162978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,512,0.003917333152559069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,256,0.004947555561860402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,512,0.003559111307064692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,768,0.006126222097211414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,256,0.004370666626426908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,128,0.004505777938498391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,2560,0.11949688858456081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,12288,0.2681093215942383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,2048,0.1404613388909234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,65536,0.2224195533328586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,4096,0.013256000147925483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,7168,0.0070675553547011475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,65536,0.13123733467525905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,1024,0.003925333420435588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,6144,0.01032977799574534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,3072,0.007039111521508958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,768,0.01218933363755544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,4096,0.008231110870838165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,128,0.011518222590287527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,2048,0.008456000023418004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,128,0.002871111035346985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,3072,0.0399235553211636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,12288,0.014734221829308404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,768,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,1536,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,7168,0.06703199942906697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,16384,0.03935733437538147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,512,0.009183111290136972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,256,0.008573333422342936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,51200,0.7626578013102213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,1536,0.012548444171746572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,3072,0.019159111711714003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,512,0.043643554051717125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,128,0.007903999752468532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,2048,0.10833777983983357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,10240,0.5704702271355523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,2560,0.0210524449745814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,3584,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,256,0.029159112109078303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,512,0.016975111431545682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,3584,0.011297778122954898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,51200,0.853150208791097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,256,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,16384,0.10071200132369995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,1024,0.004081777814361784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,1536,0.0073724447025193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,3072,0.0075804442167282104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,65536,0.8401181962754992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,2048,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,5120,0.11324711640675862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,65536,4.009092542860243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,3584,0.012317332956526013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,51200,0.03155733479393853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,512,0.004104888687531154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,3584,0.08005777994791667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,512,0.0047768888374169665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,128,0.003343111111058129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,64,0.0034222222036785553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,8192,0.018794667389657762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,3072,0.012747555143303342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,8192,0.009407111340098912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,51200,0.2744186719258626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,8192,0.014715555641386243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,512,0.01684266659948561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,4096,0.029102222787009344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,10240,0.21849332915412056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,4096,0.040271110004848905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,8192,0.011551111108726926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,65536,0.2117884423997667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,10240,0.02825777729352315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,10240,0.4215742217169867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,256,0.003216888962520493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,51200,0.1852062278323703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,1536,0.11066489087210761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,51200,0.27567021052042645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,256,0.004775111046102312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,128,0.006135111053784688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,16384,0.048531555467181735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,51200,0.9517573250664605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,2560,0.01621866722901662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,8192,0.7595591015285916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,16384,0.04513600137498644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,1536,0.005056000004212062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,1536,0.009202666580677032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,2560,0.039503110779656306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,128,0.0155102229780621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,8192,0.015256888336605497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,4096,0.014621333943472968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,12288,0.33393245273166233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,3072,0.04579200016127693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,4096,0.013411555853154926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,65536,3.757353888617622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,7168,0.19621866279178193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,64,0.019511999355422128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,256,0.009501333038012186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,3072,0.009816889133718278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,2560,0.11040266354878743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,51200,3.178793801201714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,6144,0.05690133571624756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,3072,0.010728889041476779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,8192,0.009008000294367472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,3584,0.011030221978823343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,5120,0.9569892883300781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,32,0.012796444197495779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,3584,0.14846577909257677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,51200,1.0815021726820204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,10240,0.23402844534979927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,16384,0.11182488335503472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,6144,0.1925555600060357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,2560,0.013386666774749756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,256,0.003032888803217146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,12288,0.0575431121720208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,2560,0.09525866640938653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,7168,0.01736088924937778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,2560,0.016532444291644625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,12288,0.2961440086364746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,2048,0.13421866628858778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,1024,0.006510222123728857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,512,0.004907555464241239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,768,0.010277333358923594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,64,0.0658737752172682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,8192,0.05771288606855604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,64,0.0028231110837724474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,10240,0.008583111067612966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,4096,0.00833511104186376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,3584,0.008986666798591614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,1536,0.45658312903510195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,5120,0.007035555938879649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,5120,0.06575022141138713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,3072,0.011729777687125735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,3072,0.017622222503026325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,3584,0.022970666488011677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,2048,0.01536622146765391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,1024,0.00700800038046307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,12288,0.012479999827014076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,3584,0.13294400109185114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,16384,0.18091377947065565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,256,0.0031786666562159858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,65536,0.8245546552870008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,8192,0.027079999446868896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,4096,0.013926222920417786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,7168,0.09056355555852254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,5120,0.009314667019579146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,64,0.008353778057628209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,2560,0.013202666408485837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,2560,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,12288,0.019352000620630052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,51200,0.4041431215074327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,128,0.006171555568774541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,10240,0.06055111355251736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,3584,0.011815110842386881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,5120,0.26025332344902885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,4096,0.16026843918694392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,16384,0.12415999836391872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,1536,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,6144,0.08588622013727824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,768,0.018048000004556444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,1024,0.13779732916090223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,3072,0.2942177719540066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,768,0.004361777669853634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,2048,0.0487333337465922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,65536,1.7403270933363173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,128,0.005434666656785541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,768,0.004489777816666497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,3584,1.4459964964124892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,3072,0.024243555135197107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,2560,0.10732799768447876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,2560,0.10289333264033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,8192,0.07163733243942261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,12288,2.520833757188585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,128,0.052389333645502724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,768,0.0038515557017591265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,4096,0.022257778379652236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,5120,0.01368533323208491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,7168,0.30940622753567165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,2560,0.027302222119437322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,3584,0.10911822319030762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,64,0.0026746667507621977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,5120,0.014131555954615274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,10240,0.023530667026837666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,512,0.00360622215602133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,8192,0.008937777744399177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,1536,0.00589155571328269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,1536,0.007140444384680853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,7168,0.10073333316379124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,5120,0.07165866427951388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,5120,0.009751111268997192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,5120,0.09538311428493923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,512,0.028780443800820246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,128,0.0026844443960322272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,2560,0.030841777722040813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,65536,0.664097785949707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,10240,0.21957600116729736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,10240,0.023899555206298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,4096,0.008207999997668797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,32,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,512,0.009496889180607265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,16384,0.19960444503360322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,8192,0.2004986604054769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,5120,1.573177761501736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,10240,0.372615999645657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,768,0.004353777815898259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,2560,0.027447111076778833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,3072,0.09556000100241767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,3584,0.025119110941886902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,51200,4.153667449951172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,512,0.11551200018988715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,128,0.004484444442722532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,10240,0.3823084566328261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,512,0.004355555607212914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,10240,0.023914666639433965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,10240,0.07943644126256307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,256,0.0092284447617001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,12288,0.08780178096559312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,32,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,6144,0.006688888702127669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,2560,0.04794311192300585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,2560,0.0074542222751511475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,51200,0.33830398983425564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,7168,0.6715466711256238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,8192,0.017098666893111337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,512,0.006423111177153057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,4096,0.044658667511410184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,2560,0.00655644428398874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,3072,0.09099111292097305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,1024,0.05125511023733351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,10240,0.045011556810802884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,1024,0.0038373333712418876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,32,0.004855999930037392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,16384,0.11203644010755752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,768,0.01094133324093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,512,0.05134310987260607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,2560,0.07277777459886339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,3072,0.015023999743991427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,65536,0.6271946695115831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,1536,0.007740444607204861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,64,0.028303109937243994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,6144,0.02161955502298143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,2560,0.2981342209710015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,7168,0.018294221825069852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,256,0.005720888988839255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,768,0.011197333534558615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,5120,0.06515733400980632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,51200,0.343960894478692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,1024,0.013807111316257052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,12288,1.510597334967719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,512,0.005209777918126848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,768,0.008031999899281396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,10240,7.642518785264756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,3584,0.17853866683112252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,256,0.2988559934828016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,5120,0.17259821626875135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,6144,0.015257777439223396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,12288,0.5862666765848795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,12288,0.1093591120507982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,12288,0.6608177820841471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,3072,0.0653946664598253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,16384,0.08346222506629096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,8192,0.02010311186313629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,12288,0.03081688947147793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,768,0.033907555871539645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,10240,0.03900444507598877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,64,0.0032951111594835916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,3072,0.06627821922302246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,6144,0.13035821914672852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,768,0.006243555496136348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,5120,0.04704622096485562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,2048,0.014142221874660917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,2048,0.017090666625234816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,5120,0.0064702220261096954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,512,0.1652702225579156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,2048,0.5117013189527724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,3072,0.008834666675991481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,768,0.04465511110093859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,512,0.0036755556033717263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,5120,0.03997600078582764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,12288,0.059828446971045605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,512,0.30402133199903697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,7168,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,16384,0.23216001192728677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,3584,0.10071733262803818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,768,0.009463999834325578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,2560,0.007171555525726742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,5120,0.2737084494696723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,10240,0.0624471108118693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,12288,0.013830222189426422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,2048,0.054337776369518705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,7168,0.012115555504957834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,6144,0.027020444472630818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,512,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,2048,0.8367768923441569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,12288,0.05494488941298591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,128,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,128,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,65536,0.24065777990553114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,65536,0.24357512262132433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,768,0.021264889174037512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,12288,0.23982487784491646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,6144,0.008365333080291748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,7168,0.01814133259985182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,6144,0.022502221994929846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,3072,0.020124443703227572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,8192,0.036759111616346575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,1024,0.012880889077981314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,65536,0.008992888861232335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,1536,0.042191998826132886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,2560,0.007066667079925537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,12288,0.04122666517893473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,3584,0.012055111428101858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,3072,0.01517511076397366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,2048,0.032997333341174655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,4096,0.16108444001939562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,16384,0.09419022003809611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,3072,0.011040888726711273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,16384,0.019467555814319186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,64,0.08384888701968723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,768,0.0362755556901296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,64,0.0025635556214385563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,64,0.004098666624890434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,768,0.005853333406978183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,3072,0.22809688250223795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,32,0.33165333006117076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,6144,0.26155111524793834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,4096,0.07249689102172852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,1536,0.015200000670221118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,1024,0.1360213359196981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,32,0.0027457777824666765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,3584,0.03871999846564399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,51200,0.0850951141781277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,6144,0.013439999686347114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,2048,0.011325333681371478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,7168,0.011169777976142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,8192,0.017455110947291057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,2048,0.009745777481132084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,6144,0.011018666956159802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,768,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,10240,0.48979557885064023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,128,0.009857777920034196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,1024,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,1024,0.0341360006067488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,8192,0.02650844388537937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,4096,0.038930667771233454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,6144,0.014289778139856128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,7168,0.06617422236336602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,1024,0.007876444194051955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,768,0.09648533662160237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,512,0.04276355438762241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,1024,0.19324889447953966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,32,0.0032115555885765287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,3584,0.013977777626779346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,3072,0.030123554997973975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,32,0.002879111096262932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,32,0.005706666658322017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,16384,0.41772090064154727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,10240,0.13910399542914495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,512,0.00757244477669398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,256,0.025232000483406916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,65536,0.2238559987809923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,64,0.003899555653333664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,1024,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,3072,0.010831111007266574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,7168,0.04043111205101013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,512,0.0034311111602518293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,4096,0.016999999682108562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,64,0.005684444473849402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,768,0.004379555583000183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,5120,0.02367199957370758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,256,0.004869333157936732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,128,0.04087022278043959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,5120,0.027054222093688116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,4096,0.012181333369678922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,51200,264.28488498263886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,5120,0.019474666979577806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,256,0.0047528888616296975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,51200,0.5266088909573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,3072,0.018546667363908555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,2560,0.0180711117055681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,6144,0.4410213364495172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,10240,0.12564888265397814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,1024,0.005983110931184556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,7168,0.40770665804545086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,3584,0.021537777450349595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,512,0.025571554899215698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,1536,0.02067733307679494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,2560,1.8584605322943792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,7168,0.010264000131024255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,512,1.5474276012844508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,4096,0.012792000340090858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,1536,0.0380586650636461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,6144,0.2897324562072754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,51200,0.5494169129265679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,128,0.003549333247873518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,3072,0.05860800213283963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,768,0.019480889042218525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,6144,0.010784000158309937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,512,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,6144,0.016554666890038382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,10240,4.688912709554036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,65536,0.13292533821529812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,3584,0.00905600024594201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,768,0.012117333710193634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,2560,0.01054044481780794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,256,0.003556444413132138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,3584,0.03574400146802267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,6144,0.03085955646302965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,6144,0.18957244025336373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,16384,0.021184000704023574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,512,0.0799057748582628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,4096,0.0069573331210348345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,512,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,1536,0.005481777919663324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,1024,0.06432088878419664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,65536,0.07855021953582764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,128,0.005063111169470681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,16384,0.23404177029927573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,3584,0.48936621348063153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,3584,0.26212888293796116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,3072,0.011755555868148804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,7168,0.31371911366780597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,51200,1.3427111307779949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,4096,0.08074311415354411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,65536,0.301329771677653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,128,0.011160888605647616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,7168,0.19368267059326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,2560,0.02409066590997908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,2560,0.014268444644080268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,51200,0.6246222390068902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,16384,0.22777954737345377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,5120,0.012505778008037143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,2048,0.011994666523403592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,2048,0.01036711119943195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,10240,16.37604437934028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,32,0.018255111243989732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,16384,0.05533155467775133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,3072,0.01200711147652732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,128,0.005176888985766305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,64,0.05531911055246989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,16384,0.5978826416863335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,32,0.0029066666546795103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,64,0.004078222231732475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,32,0.005767111149099138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,3072,0.05150933398140801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,2048,0.013745778136783175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,7168,0.04738577869203356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,4096,0.06303733587265015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,8192,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,128,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,512,0.004550222307443619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,7168,0.4552666876051161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,5120,0.8395599789089627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,64,0.00832711077398724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,768,0.03226133187611898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,2048,0.017456000049908955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,2048,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,5120,0.016912000046836007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,32,0.005205333232879639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,256,0.4910737673441569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,65536,0.009595555563767752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,768,0.003540444291300244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,32,0.008387555678685507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,5120,0.30083200666639537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,3072,0.007296889192528195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,128,0.003670222229427762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,2560,0.029117332564459905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,8192,0.025179555018742878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,2048,0.007261333366235097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,7168,0.13156267007191977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,3584,0.008036444584528605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,3584,0.023539554741647508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,10240,0.013611555927329592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,5120,0.38477333386739093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,5120,0.11318755149841309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,1536,0.21931466791364881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,256,0.005872000008821487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,2560,0.01755644381046295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,128,0.0033048888047536216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,3072,7.85736083984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,51200,0.13918488555484348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,3072,0.007721777591440413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,512,0.005017777697907554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,5120,0.013716444373130798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,7168,0.045575112104415894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,64,0.012643555800120035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,4096,0.07424533367156982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,5120,0.03581066595183478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,16384,1.1854542626274955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,1024,0.013195555243227216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,3072,0.02843466732237074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,12288,0.3973866568671332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,2048,0.01921866668595208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,3584,0.013245333400037555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,16384,0.11072355508804321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,6144,0.0101724440852801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,32,0.002728888971938027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,4096,0.0069573331210348345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,3072,0.1340577734841241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,6144,0.016730666160583496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,1536,0.03639466563860575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,51200,0.03202222122086419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,128,0.0052782222628593445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,128,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,8192,0.12261244985792373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,6144,0.24502489301893446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,8192,0.48426400290595156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,768,0.009865778187910715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,51200,0.11627377404106988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,51200,2.256411658393012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,65536,0.25641687711079914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,2048,0.0070168889231152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,12288,0.0196088883611891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,51200,3.676324632432726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,65536,7.757910834418403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,1536,0.0047102222839991255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,16384,0.7891502380371094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,3584,0.009161777794361115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,5120,0.07246488995022245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,64,0.0057813334796163775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,3072,1.0457671483357747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,2560,0.07646222247017755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,12288,0.9295519722832574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,4096,0.025645333859655593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,5120,0.03706222110324436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,7168,0.11777688397301568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,12288,0.06693955262502034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,768,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,1024,0.018225777480337355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,512,0.08480266729990642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,128,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,1536,0.008590222232871586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,512,0.0032479998966058097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,32,0.003244444520937072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,16384,1.085770712958442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,512,0.33674221568637425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,768,0.009764444496896531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,128,0.0058000000814596815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,3584,0.09404089053471883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,7168,0.06553244590759277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,2048,0.007521777517265744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,5120,0.11709866258833145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,5120,0.011610666910807291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,32,0.040977779361936785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,3072,0.01646488904953003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,16384,0.18868533770243326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,16384,0.04317155480384827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,256,0.0032640000184377036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,10240,0.07178933090633817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,10240,0.0243493335114585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,1024,0.03774844606717428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,1024,0.010913777682516309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,65536,0.9121546215481229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,1024,0.04467199908362495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,2048,0.38048444853888613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,1536,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,51200,0.05528710948096382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,256,0.005408888889683618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,1024,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,64,0.00349866681628757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,3584,0.009487999810112847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,32,0.0030657777355776895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,512,0.005311111195219888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,512,0.005821333163314396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,2560,0.010500444306267632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,12288,0.026943110757403906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,3584,0.016582222448454965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,12288,0.019313777486483257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,3584,0.05012977785534329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,128,0.006581333362393909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,6144,0.012306667036480375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,7168,3.4848880767822266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,16384,0.08462222417195638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,768,0.00795288880666097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,4096,0.08792089091406928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,2560,0.04345866706636217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,5120,0.050063110060162015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,512,0.005767999837795894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,3584,0.011568889021873474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,16384,1.7982745700412328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,10240,0.009436444275909001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,12288,0.03480888737572564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,5120,0.06820444266001384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,1024,0.005712000032265981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,3072,0.010727110836240979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,3584,0.02816533380084568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,1024,0.3816080093383789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,5120,0.013807999591032663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,64,0.007400888535711501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,256,0.00924355536699295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,2560,0.05927466683917575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,4096,0.03616266780429416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,3584,0.02217866645918952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,2048,0.4874640040927463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,5120,0.012121777567598553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,6144,0.026514665948020086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,10240,0.1263200044631958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,2048,0.1786808835135566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,256,0.02050133380624983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,64,0.0031031111462248694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,1536,0.0496711101796892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,64,0.040848887628979154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,8192,0.08887022071414524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,6144,0.3502355681525336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,10240,0.03358577688535055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,3072,0.20125956005520293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,128,0.05807022253672282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,5120,0.4837315347459581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,1536,0.012320888539155325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,256,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,2048,0.009728888670603434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,3072,0.011271110839313932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,51200,2.2290435367160373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,3584,0.010763555765151978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,32,0.006969778074158563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,65536,3.2441654205322266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,768,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,512,0.037616888682047524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,256,0.003855111284388436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,256,0.0030986666679382324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,12288,0.07344977723227607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,7168,0.010604444477293225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,4096,1.7599431143866644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,1536,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,51200,0.08568977647357517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,2560,0.008496888809733922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,1536,0.007163555257850223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,12288,0.07101244396633573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,51200,0.20248888598548043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,16384,0.04361244373851352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,2048,0.014474666780895658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,2048,0.6751004325018989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,1536,0.034789333740870156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,2048,0.0068408888247278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,2048,0.011981333295504252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,1024,0.03457599878311157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,4096,0.6288755734761556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,256,0.02234755622016059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,4096,0.0941386686431037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,12288,0.010780444575680627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,768,0.005365333209435145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,1024,0.008918222453859117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,7168,0.021495999561415777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,2560,0.014887111054526435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,16384,0.02404533326625824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,6144,0.45957422256469727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,8192,0.031685332457224526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,6144,1.2557528813680012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,2048,0.008647110727098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,1024,0.00647555540005366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,64,0.003302222324742211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,2048,0.06999822457631429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,768,0.005928888916969299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,2048,0.007425777614116669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,65536,14.989193386501737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,8192,0.039910223748948835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,16384,0.02427822185887231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,16384,0.01977688901954227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,64,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,7168,0.03805244300100539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,65536,1.548087967766656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,65536,0.3958124319712321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,256,0.004239999999602635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,3584,0.17216622829437256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,1536,0.015600888265503777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,6144,0.010189333723651038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,256,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,65536,0.03988800115055508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,8192,0.016752888758977253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,768,0.005629333357016246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,5120,0.009335111412737105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,32,0.005161777966552311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,1536,0.01145066652033064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,256,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,6144,0.013285333083735572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,51200,0.0244186669588089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,1024,0.003992889076471329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,16384,0.07541688945558336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,8192,0.020389333367347717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,768,0.010027555955780877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,128,0.009149333669079674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,2560,0.007631111476156447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,7168,0.006123555617200003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,3072,2.325094223022461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,128,0.005998222364319696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,12288,0.30859822697109646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,2048,0.07327022155125935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,4096,0.13020266426934138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,2560,0.009935999910036722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,4096,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,5120,0.014499555031458536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,256,0.010371555884679159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,5120,0.07310399744245741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,512,0.003362666608558761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,4096,0.18578755855560303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,128,0.040008889304267034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,1536,0.08552444643444485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,51200,0.04138488901986016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,32,0.0070462218589252895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,1536,0.10689955287509495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,5120,0.009352888498041365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,8192,0.44155022833082413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,512,0.003946666502290302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,12288,0.4077920118967692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,3584,0.00795911086930169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,5120,1.8226952022976344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,2560,0.013380444712109037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,2560,0.04762666755252414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,768,0.006314666734801398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,3072,0.019056000643306308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,65536,0.06855377886030409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,128,0.005699555493063397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,128,0.004972444640265571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,7168,0.04197866717974345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,768,0.3174915578630236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,2560,0.012469333079126147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,16384,0.13742578029632568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,4096,0.01369955556260215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,1536,0.01163200040658315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,8192,0.0522346662150489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,10240,0.019056888090239633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,3584,0.15219288402133518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,1024,0.06527200010087755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,2560,0.12348532676696777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,16384,0.10457422335942586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,3072,0.008283555507659912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,10240,0.029767112599478826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,65536,1.0254347059461806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,3072,0.016861332787407767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,1536,0.00665244460105896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,12288,1.8602080874972875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,6144,0.17280977302127412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,3072,0.01900533338387807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,3072,0.010963555839326648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,4096,1.0277218288845487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,3584,0.7381528748406304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,64,0.00654666663871871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,3072,0.013924444715181986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,6144,0.5157253477308485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,3072,0.056932442718082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,5120,0.02404622236887614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,7168,0.04134044382307265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,16384,0.02281155520015293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,768,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,32,0.003850666599141227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,5120,0.04598399996757507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,1024,0.026017778449588354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,10240,0.3025493356916639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,256,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,512,0.008750222623348236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,65536,1.2508195241292317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,2048,0.00813244448767768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,7168,2.656466590033637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,8192,0.07188710901472303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,3584,0.009438222481144799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,16384,0.1174515618218316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,10240,0.029147555430730183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,1024,0.06341599755816989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,64,0.012527111503813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,32,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,2560,0.19399110476175943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,2048,0.007039111521508958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,6144,0.16792799366845024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,2560,0.014534221755133735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,128,0.020613332589467365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,5120,0.010666666759385003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,256,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,128,0.004312889029582341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,1024,0.010405333505736457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,3584,0.043763554758495755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,4096,0.026935999592145283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,1536,0.014271999398867289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,12288,0.0839315586619907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,1024,0.021447999609841242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,1536,0.05840088923772176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,2048,0.008549333446555668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,128,0.04527911212709215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,8192,0.012220444778601328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,3072,0.03772355450524224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,12288,0.16588089201185438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,8192,27.56561279296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,8192,0.03846222162246704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,64,0.031503111124038696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,768,0.030594666798909504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,768,0.004651555584536659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,128,0.006236444330877728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,768,0.01275466630856196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,5120,0.03881866733233134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,3072,0.09222222036785549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,8192,0.13026844130622015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,65536,0.17306221856011283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,8192,0.25106578403049046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,65536,4.16578843858507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,16384,0.20104444026947021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,512,0.13927911387549505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,768,0.08895466725031535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,2560,1.0718648698594835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,768,0.007312888900438945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,32,0.012768000364303589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,3584,0.04838133189413282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,768,0.009962666365835402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,1024,0.009161777794361115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,10240,0.1712568865882026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,3584,0.03852799865934584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,7168,0.06352621979183622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,1024,0.009916444619496664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,4096,0.12538577450646296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,128,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,6144,0.6885280079311795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,3072,0.019483556350072224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,1536,0.02886311213175456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,512,0.021344888541433547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,128,0.011959111524952782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,1024,0.11188799805111355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,7168,0.01961955593691932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,512,0.07914044459660848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,6144,2.6282391018337674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,5120,0.017697778013017442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,768,0.010431999961535135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,128,0.005468444277842839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,2048,0.23891377449035645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,128,0.07231644127104017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,6144,0.012893333203262754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,32,0.003060444361633725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,64,0.003735110991530948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,12288,0.02516799999607934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,4096,0.017986666825082567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,128,0.016880000631014507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,1536,0.07487289110819499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,5120,0.006647999915811751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,2560,0.011686222420798408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,16384,0.02405866649415758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,2048,0.010079111489984725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,10240,0.01754133403301239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,51200,0.13455288940005833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,5120,0.5344844394259983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,10240,0.14777244461907282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,6144,0.00980888886584176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,1536,0.02823644545343187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,1536,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,12288,0.030369778474171955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,51200,0.2288595570458306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,1536,0.016511999898486666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,768,0.0041866666740841335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,128,0.0028817777832349143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,1024,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,768,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,10240,0.09739022122489081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,64,0.028078221612506445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,2048,0.011157333023018308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,4096,0.01036800030204985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,3584,0.09640533394283718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,3072,0.07466044690873888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,3072,0.03550844391187032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,512,0.013835555149449242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,65536,3.2452996571858725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,64,0.13210311200883654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,65536,20.03445095486111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,2048,0.07804622252782185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,16384,0.2227484385172526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,256,0.03340266810523139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,512,0.005687110953860813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,128,0.015211555692884656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,1536,0.023562666442659166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,64,0.011058666639857821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,1536,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,768,0.008346666892369589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,7168,0.010475555227862464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,2560,0.009628444082207149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,3584,0.0865866674317254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,1024,0.004048888882001241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,7168,0.030607998371124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,1024,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,6144,0.013040000365840064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,2048,0.011552888486120435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,4096,0.012159999873903064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,1024,0.004577777865860197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,1024,0.00424888895617591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,16384,0.5278017785814073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,16384,0.015817777978049386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,512,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,10240,0.27408266067504883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,1024,0.007152000235186682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,12288,0.3953191174401178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,5120,0.022392888863881428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,51200,0.01333688861793942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,51200,0.8068186442057291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,2048,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,3072,0.025415110919210646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,1536,0.028212444649802312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,128,0.0028853333658642242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,10240,0.15432977676391602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,16384,0.03703822361098395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,768,0.007654222349325816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,65536,1.268952899509006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,7168,0.23163021935356987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,768,0.004399111287461387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,128,0.009889778163697984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,2048,0.018876444962289598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,4096,0.3180977768368191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,51200,28.459238688151043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,51200,37.03572252061632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,65536,0.10222755538092719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,16384,0.042099555333455406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,32,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,16384,5.789696163601345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,1024,0.010713777608341642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,1024,0.009315555294354757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,65536,0.08643555641174316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,128,0.12061066097683376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,51200,0.05618933174345228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,768,0.006170666466156642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,12288,0.017064889272054035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,1536,0.005639111002286275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,2048,0.007084444165229797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,12288,0.07216266791025798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,128,0.004368888835112254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,256,0.01700266698996226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,64,0.00518488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,32,0.02085866696304745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,12288,0.03915466533766852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,12288,0.0184897780418396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,128,0.004383999854326248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,64,0.006445333361625671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,2048,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,6144,0.0884551140997145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,512,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,2560,0.018329777651362948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,128,0.019655999210145738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,16384,2.9175768958197703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,7168,0.11420267158084446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,512,0.0041662222809261745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,5120,0.02379377848572201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,51200,1.7789190080430772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,16384,0.04140888982348972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,3584,0.3368391195933024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,2560,0.07769244247012667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,768,0.005946666830115848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,6144,0.02252799934811062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,51200,12.03798082139757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,16384,0.04142399960094028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,2560,0.008756444685988957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,16384,3.0436888800727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,7168,0.17325511243608263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,10240,0.023842665884229872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,2048,0.04655999938646952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,3584,0.0211173329088423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,1024,0.006491555521885554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,1024,0.009163555171754625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,64,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,256,0.011206222077210745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,4096,0.012102222277058495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,65536,0.06838488578796387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,3584,0.14908622370825872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,3072,0.0544177790482839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,128,0.005020444591840108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,768,0.008822222550710043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,256,0.007032000356250339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,7168,0.05968088573879666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,256,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,3584,0.06976622343063354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,3072,0.07109510898590088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,4096,0.039348443349202476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,2560,0.013790222505728403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,256,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,8192,0.2503440115186903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,10240,0.17075110806359184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,128,0.003428444473279847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,1024,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,7168,0.006314666734801398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,10240,0.01648355523745219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,5120,0.0354728897412618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,16384,0.09313333034515381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,1536,0.006020444548792309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,2048,0.017246221502621967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,2048,0.026791110634803772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,10240,4.22342766655816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,7168,0.016187555260128446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,10240,0.026371555195914373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,2048,0.03360889024204678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,16384,0.05314755439758301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,3072,0.06086844205856323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,12288,0.09729688697391087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,12288,0.030552887254291113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,3072,0.017145777742067974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,12288,0.037752889924579196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,6144,0.006361777583758037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,16384,2.65248171488444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,64,0.12693866093953451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,8192,0.027074666486846075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,10240,0.02403555479314592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,768,0.006065777606434292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,6144,0.9194293551974826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,3584,0.014947555131382413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,768,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,768,0.0660159985224406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,2048,1.3589528401692708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,768,0.03163289030392965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,8192,0.293034659491645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,10240,0.415559106402927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,128,0.004307555655638377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,2048,0.008480889101823172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,128,0.005568000177542369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,6144,0.2781733406914605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,2560,0.018505778577592637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,1024,0.012864888542228274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,2560,0.06271200047598945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,16384,0.07806399795744154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,32,0.00527555536892679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,6144,0.005978666659858491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,1536,0.009581333233250512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,1536,0.004719999929269155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,65536,0.24620532989501953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,3072,0.6016915639241537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,5120,0.03224888775083754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,8192,1.4661359786987305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,256,0.006627555522653792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,16384,0.4037004576789008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,128,0.004281777888536453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,16384,0.7174213197496203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,12288,0.49771467844645184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,1536,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,128,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,2048,0.010263111028406356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,3072,2.051004409790039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,2048,0.024688889582951862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,128,0.0030986666679382324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,8192,0.12490755981869167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,1536,0.013733333183659447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,16384,0.023768888579474554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,65536,0.1463128858142429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,1024,0.006219555520349079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,4096,0.08902311325073242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,64,0.009522666533788046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,128,0.004596444467703502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,4096,0.10999377568562825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,5120,0.013914666242069669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,3072,0.040412445863087974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,2560,0.01613333324591319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,12288,0.02184533410602146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,1536,0.053857776853773326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,768,0.011494222614500256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,6144,0.025776000486479864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,128,0.06617599725723267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,64,0.002691555561290847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,3584,0.11315911346011692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,16384,0.24098398950364855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,64,0.0027502222607533136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,65536,0.8764035436842176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,8192,0.04052622119585673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,4096,0.016693333784739178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,256,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,16384,0.2258168856302897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,4096,0.05615644322501289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,32,0.002642666714058982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,128,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,768,0.02514133354028066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,64,0.0029271110478374693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,2048,0.0734835531976488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,768,0.00870488915178511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,32,0.04056444433000352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,3584,0.17670667171478271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,512,0.04673244555791219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,1536,0.083697775999705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,768,0.019176888797018263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,128,0.00544088871942626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,7168,0.010880000061459012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,32,0.009594666461149851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,51200,0.09805866744783188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,12288,0.15149511231316462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,5120,0.011164444188276926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,8192,0.022849778334299724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,2048,0.25221867031521267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,1024,0.006907555378145642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,5120,0.02308533257908291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,256,0.04682844546106127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,4096,0.010202666951550378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,65536,1.8349689907497828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,6144,2.7536275651719837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,1536,0.0075297777851422625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,51200,0.2625031206342909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,512,0.2797413402133518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,3584,0.007864000068770515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,12288,0.3927946620517307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,51200,0.05493333273463779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,10240,0.016629333297411602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,2560,0.1347564458847046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,65536,0.5364258024427626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,1536,0.06421244144439697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,256,0.005254222287072076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,7168,0.11903822422027588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,65536,0.08499733606974284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,3584,0.017436444759368896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,256,0.013181333740552267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,4096,0.01074044406414032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,12288,0.08210488822725084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,8192,0.03526577684614394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,768,0.3431688944498698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,8192,0.01494311127397749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,8192,0.40082044071621364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,51200,0.06494310829374525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,3072,0.01605599953068627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,64,0.0038977778620190094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,4096,0.013902222116788229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,2560,0.014720888601409065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,3584,0.013418667018413544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,2560,0.019223110543357003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,2560,0.02932888931698269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,7168,0.012198222180207571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,512,0.0047004446387290955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,51200,0.3269511063893636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,65536,0.4006497859954834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,10240,0.04810577630996704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,4096,0.042111999458736844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,768,0.13365688588884142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,768,0.006019555446174409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,128,0.003269333392381668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,512,0.003535110917356279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,7168,0.40737777286105686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,65536,0.008184889124499427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,512,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,128,0.015978667471143935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,7168,0.235938655005561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,5120,0.055833776791890465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,512,0.003592888928121991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,256,0.05699555741416084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,8192,0.02036266691154904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,65536,3.3731945885552297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,1024,0.021386666430367365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,3584,0.10956622494591607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,65536,1.7571919759114583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,4096,0.013669333524174161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,65536,0.19510755274030897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,512,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,51200,0.06549421946207683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,16384,0.07450844181908502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,768,0.05672978030310737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,3072,0.015990222493807476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,3072,0.007247111035717859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,3584,0.028593778610229492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,10240,0.058515555328793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,32,0.009911999934249455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,1536,0.006261333409282897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,2560,0.048892445034450956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,512,0.02762311034732395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,2048,0.045555555158191256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,64,0.010032888915803697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,7168,0.015245333313941956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,3584,0.10195999675326878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,256,0.010791999598344168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,7168,0.013613333304723104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,2048,0.010686222049925061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,8192,0.15482932991451687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,12288,0.007583111524581909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,5120,0.00945688866906696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,1024,0.03657866517702738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,2048,0.011131555669837527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,512,0.0038577777643998465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,4096,0.14970755577087402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,65536,0.17993511093987358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,5120,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,4096,0.008947555389669206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,768,0.003583999971548716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,8192,0.09627022345860799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,3584,0.01348622226052814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,5120,0.009409777820110321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,65536,0.19004800584581164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,512,0.00518488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,51200,0.5023431248135036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,2560,0.007320889168315464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,32,0.0058328890138202244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,5120,0.02145244512293074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,64,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,1536,0.02144177754720052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,3072,0.012078222301271228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,65536,0.24663199318779838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,768,0.0058337777025169795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,3072,0.09340266386667888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,8192,0.021495999561415777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,1536,0.007152000235186682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,10240,0.7455786599053277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,128,0.004906666775544484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,5120,0.06228622463014391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,1024,0.004571555389298333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,128,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,2048,0.008380444513426887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,2048,0.08800799979103936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,128,0.017309332887331646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,16384,4.832259707980685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,3584,0.14659555753072104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,10240,0.12001866764492458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,8192,0.015007111761305066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,5120,0.03732266690995958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,8192,3.3848372565375437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,10240,0.3445555633968777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,3584,0.008582221964995066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,16384,0.1648062202665541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,64,0.00463733325401942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,4096,0.02941688895225525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,64,0.011596444580290051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,64,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,2560,0.015356444650226168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,32,0.005200888961553574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,12288,0.07667555411656697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,1536,0.010636444720957013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,3584,0.06286311149597168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,4096,0.010129777921570672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,5120,0.06676177846060859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,3072,0.17364711231655547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,4096,0.008618666893906063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,128,0.005191111316283544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,2048,0.12808088461558023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,3584,0.07642133368386163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,768,0.005449777675999536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,64,0.005817777580685086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,6144,0.027691556347741023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,1024,0.060311112138960093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,512,0.005630222045713001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,8192,0.007234666910436418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,2560,3.78515625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,7168,1.7633999718560114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,6144,0.05724710888332791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,6144,0.07909422450595431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,512,0.005072000126043956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,2048,0.006630222416586346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,12288,0.4009688960181342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,65536,7.411735534667969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,65536,0.08435110913382636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,10240,0.11065866549809773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,65536,15.163497924804688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,3072,0.012510221865442065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,6144,0.043531556924184166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,64,0.003084444544381566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,128,0.014581332604090372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,3584,1.3710098266601562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,256,0.004006222304370669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,5120,0.014228444960382251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,128,0.06313955783843994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,51200,0.07108622127109103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,1536,0.03670933180385166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,1536,0.06202222241295708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,16384,0.006248000181383557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,128,0.0069102222720781965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,12288,0.3889724413553874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,512,0.04398666818936666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,7168,0.011531555818186866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,256,0.0047395556337303585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,65536,0.10914933019214207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,32,0.0047848886913723415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,128,0.02859555681546529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,128,0.005164444446563721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,1536,0.03966133462058173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,4096,0.6244559817843968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,51200,0.16502400239308676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,2048,0.00683822234471639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,1024,0.04946310983763801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,3072,0.09744711054695977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,32,0.061576002173953585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,128,0.003910222401221593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,3072,0.02200711104604933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,256,0.003458666718668408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,3072,0.007559110720952352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,2560,0.14881866508060032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,16384,0.04428799947102865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,12288,0.02682311170630985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,10240,0.05535111162397596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,3584,0.006664888726340399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,2560,0.01499555508295695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,7168,0.020083554916911654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,3584,0.015394666128688388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,10240,0.03901688920127021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,64,0.002809777855873108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,8192,0.3089360131157769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,768,0.006528888725572162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,3584,0.014096000128322177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,3584,0.02945599953333537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,65536,0.015204444527626038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,3072,0.01943466729587979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,51200,0.15396532747480604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,3072,0.09034133619732326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,4096,0.014516444669829475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,256,0.00527555536892679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,2048,0.0070426662762959795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,7168,0.21748267279730904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,1536,0.012359110845459832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,5120,0.02403644389576382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,256,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,5120,0.03052088949415419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,51200,0.015034667319721647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,10240,0.04924177792337206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,6144,0.028765334023369685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,3584,1.3494897418551977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,128,0.014920888675583733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,2048,0.02528000043498145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,7168,5.906213124593099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,3072,0.014008000493049622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,64,0.012635555532243518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,12288,0.02293155590693156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,768,0.008371555142932469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,4096,0.07826399803161621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,4096,0.01276088919904497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,10240,0.17531111505296496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,2560,0.12006489435831706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,10240,0.27141422695583767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,3072,0.014334221680959066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,8192,0.035255998373031616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,1024,0.022314666046036616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,128,0.007893333004580604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,10240,0.11590311262342666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,3584,0.019015999303923715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,1536,0.006076444354322221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,6144,0.19700177510579428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,6144,0.09969866938061184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,2560,0.008268444074524773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,1536,0.10698222451739842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,3584,0.010055999788973067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,768,0.018844443890783522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,10240,0.5325422286987305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,2048,0.009759110709031424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,51200,0.13821332984500462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,6144,0.016460445192125108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,768,0.020788444413079154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,65536,0.11245066589779323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,128,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,51200,0.14495200581020778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,4096,0.008577778107590145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,3072,0.01811377704143524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,32,0.021151999632517498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,51200,6.643353780110677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,1024,0.08853777911927964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,10240,0.19341688685946992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,16384,0.1849679946899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,7168,2.6508621639675565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,512,0.004944889081848992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,256,0.030894223186704848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,64,0.006955555743641323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,16384,0.026086222794320848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,5120,0.03883822096718682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,51200,1.4456133312649195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,7168,0.0366960002316369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,16384,0.03143822153409322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,768,0.01865511139233907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,10240,1.3209831449720595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,128,0.02028622229894002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,128,0.02444533341460758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,51200,0.804167111714681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,32,0.02120977805720435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,1024,0.004373333520359463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,10240,0.2668844328986274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,2048,0.01944266590807173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,5120,0.015104000767072042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,64,0.0029235554652081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,5120,0.012811555630630918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,768,0.008046222229798635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,3584,0.08552710877524482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,51200,1.6236897574530706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,16384,3.7296799553765188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,5120,0.008748444418112436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,32,0.008365333080291748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,1024,0.0074746666683091064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,3584,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,768,0.013354666531085968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,256,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,16384,0.07448799742592706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,2048,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,6144,0.03372711274358962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,128,0.013951111170980664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,7168,0.11050933599472046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,2048,0.011266666981909009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,64,0.027950220637851294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,10240,0.029727998707029555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,6144,0.03467911150720384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,32,0.0026337777574857077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,4096,0.013555555707878537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,768,0.003967111309369405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,3072,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,1536,0.06791288985146417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,51200,32.36017523871528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,12288,0.14011377758449978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,12288,0.031453334622912936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,7168,0.02108088963561588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,8192,0.10353422164916992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,256,0.003378666730390655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,32,0.01984355515903897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,768,0.004552000098758274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,1536,0.012917333179050021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,8192,0.010105777945783403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,5120,0.08973333570692275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,768,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,768,0.023919110496838886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,2560,0.04421866602367825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,4096,0.08500888612535264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,3072,0.007374222079912822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,2048,0.007961778177155389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,51200,4.309003618028429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,6144,0.02033333314789666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,3072,0.015601777368121676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,10240,0.531617800394694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,3584,0.14951377444797093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,51200,0.00906577789121204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,2048,0.009662222531106737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,3584,0.015863999724388123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,8192,0.2161360051896837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,7168,0.038147555457221136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,128,0.16499111387464735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,12288,0.01940088967482249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,2560,0.05154044429461161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,8192,0.016579555140601266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,256,0.00648266656531228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,16384,1.6516417397393122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,256,0.035291555855009295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,1024,0.007237333390447829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,2560,0.012310222619109683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,2048,0.010575999816258749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,5120,0.009288888838556077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,128,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,3584,0.006804444309737947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,768,0.06947377655241224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,512,0.11637422773573135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,512,0.01018488903840383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,51200,0.11086044708887736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,12288,0.13695200284322104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,7168,0.22448266877068412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,6144,0.5111021995544434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,16384,0.12602310710483128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,51200,0.22192533810933432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,1536,0.016303999556435477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,2560,0.007866666548781926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,16384,0.053485333919525146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,2048,0.022080888350804646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,7168,0.22868090205722383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,2048,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,12288,1.1746062172783744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,6144,0.14206666416592068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,2560,1.0141999986436632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,128,0.009325332939624786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,2560,0.010307555397351583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,3072,0.033003555403815374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,3072,0.009559111462699043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,256,0.00812533332241906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,51200,0.717650678422716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,3584,0.02873599860403273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,51200,0.3947582244873047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,3584,0.012667555775907306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,128,0.0037155557009908888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,2560,0.08780711226993138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,6144,0.020756444997257657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,65536,0.11575644546084934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,512,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,128,0.004335999902751711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,1024,0.01571555601225959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,3072,0.006879111131032308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,512,0.013720889058378009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,12288,0.06404444244172838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,6144,0.010827555424637265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,5120,0.056236445903778076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,32,0.012142221960756512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,512,0.006339555399285422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,1024,0.0047591109242704184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,16384,0.12415377298990886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,1536,0.015054222610261707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,16384,0.026476444469557867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,2560,0.041518221298853554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,7168,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,512,0.008012444608741337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,65536,0.7751333448621961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,3584,0.01106844428512785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,65536,2.904058668348524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,5120,0.010354667074150508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,2048,0.009738666315873465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,12288,0.4171191056569417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,7168,0.1559280024634467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,5120,4.788309309217665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,128,0.012110222544935016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,3072,0.03313600023587545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,64,0.0035031110876136353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,16384,0.03757066528002421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,64,0.14138132996029323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,1024,0.016028443972269695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,512,0.03543555405404832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,7168,0.013608000344700284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,32,0.0058328890138202244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,10240,0.060758219824896924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,2048,0.020576000213623047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,10240,0.24459287855360243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,32,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,7168,0.012703999876976013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,16384,5.801356421576606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,16384,0.12045155631171332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,256,0.007688889073001013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,64,0.0041546668443414895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,7168,0.020768000019921195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,512,0.004579555657174853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,8192,0.06588533189561632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,3072,0.009082666701740688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,5120,0.011624889241324531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,2048,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,512,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,12288,0.01780177818404304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,5120,0.07811555597517225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,8192,0.05197777681880527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,4096,0.009597332941161262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,3072,0.00906844437122345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,256,0.005249777601824866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,1024,0.006917333437336816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,2048,0.08082489172617595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,2560,0.19584443834092882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,4096,0.010763555765151978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,6144,0.013002666334311167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,7168,0.7873999807569715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,768,0.03311466508441501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,8192,0.4793715476989746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,3584,0.007299555672539606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,6144,0.017851556340853374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,7168,0.2895875506930881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,10240,0.031112889448801678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,65536,0.14266578356424967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,1536,0.08380799823337132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,512,0.17874933613671196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,7168,0.4304808775583903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,65536,0.6141520076327854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,12288,0.11202843983968098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,16384,0.13390221860673693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,256,0.003640888879696528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,6144,0.30057067341274685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,1536,0.012469333079126147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,2560,0.007744000189834171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,2048,0.0075431110130416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,5120,0.012019555601808759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,3584,0.03615555498335097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,8192,0.04403644469049242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,128,0.00830488900343577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,256,0.02979733215437995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,768,0.01883288886811998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,2048,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,51200,0.07443644603093465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,6144,0.07247200277116564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,65536,0.18320355150434706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,2560,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,256,0.008591111335489485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,51200,0.8994933234320747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,32,0.025972444150182936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,7168,0.01831644442346361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,16384,0.04316088888380262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,512,0.06268266836802165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,8192,0.031002667215135362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,6144,0.21207643879784477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,128,0.0053680001033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,512,0.015610666738616096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,12288,0.06416089004940458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,5120,0.02930844492382473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,7168,0.006109333286682765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,51200,0.974930657280816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,1536,0.010247999595271217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,768,0.02586222191651662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,3584,0.09453688727484809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,1024,0.01092355532778634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,51200,0.08849866522683038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,4096,0.015523556205961438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,10240,0.019699555304315355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,3584,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,3072,0.09621244668960571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,7168,0.4731013509962294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,32,0.003240888938307762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,128,0.017550221747822232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,7168,0.032551109790802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,2048,0.08576711018880208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,1024,0.011736888852384357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,4096,0.010326222413116032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,12288,0.1831137736638387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,51200,4.667721642388238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,5120,0.00674044465025266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,12288,0.2929955588446723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,7168,0.012161778079138862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,10240,0.07719822062386407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,12288,0.1644408835305108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,5120,0.015804444750150044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,6144,0.22838756773206922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,1024,0.038972444004482694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,6144,0.2399137814839681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,512,0.0333591103553772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,51200,2.2533929612901473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,3584,0.013153778182135688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,768,0.0209840006298489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,5120,0.014577777849303352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,7168,0.02377333409256405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,2048,0.16982221603393555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,12288,0.5443101988898383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,1024,0.011697777443461947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,1536,0.06004799736870659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,1024,0.021390222840838965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,512,0.01906222270594703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,768,0.06803733110427856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,1024,0.004259555704063839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,51200,1.4850524266560872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,1536,0.037977778249316745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,1536,0.008261333737108443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,64,0.0031271111220121384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,7168,0.02162844439347585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,256,0.0035822221802340615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,256,0.011490667031870948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,51200,0.1316159963607788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,5120,0.011332444846630096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,6144,0.024313777685165405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,1024,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,51200,0.031530665026770696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,16384,0.03258133265707228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,51200,0.8297671212090386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,10240,0.03214577833811442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,8192,0.08222311072879367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,1024,0.0058337777025169795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,7168,0.20846045017242432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,16384,0.6341590881347656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,1536,0.018561777141359117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,65536,2.163238313462999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,2560,0.009295110901196798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,128,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,2560,0.150160895453559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,51200,3.6546062893337674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,12288,0.05485066771507263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,64,0.017312000195185345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,2560,0.01178311142656538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,5120,0.06318133407168917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,32,0.0029306666304667792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,768,0.0047351109484831495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,1536,0.065920889377594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,1536,0.012071111136012606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,5120,0.5571644571092393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,16384,0.04138311081462436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,2048,0.010401777923107147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,3584,0.06115733252631294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,512,0.006081777728266186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,768,0.021149333980348375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,12288,0.02533066603872511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,5120,0.015295111470752291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,12288,3.5660620795355906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,4096,0.012706666356987424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,12288,0.02622844444380866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,3584,0.021511110994550917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,64,0.025751110580232408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,16384,0.035210665729310774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,5120,2.2826781802707248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,1024,0.07005955775578816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,1024,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,6144,0.25561510192023384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,1024,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,1536,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,6144,0.13096888860066733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,1536,0.08661066823535496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,6144,0.0210826661851671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,51200,1.1770586437649198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,16384,0.025174222058720056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,128,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,2048,0.006604444649484422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,8192,0.039919111463758684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,2560,0.019855111837387085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,12288,0.10555733574761285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,65536,0.8343066639370389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,256,0.00701422244310379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,7168,0.03715111149681939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,7168,2.4776808420817056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,6144,0.03229422370592753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,256,0.01740533279048072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,2560,0.052015999952952065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,768,0.010961777634090848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,7168,0.02128977742460039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,3584,0.2705502245161268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,256,0.007816000117195977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,8192,0.15257067150539821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,32,0.00812177773978975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,51200,2.9434106614854603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,10240,0.04184800055291918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,4096,0.1491946644253201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,768,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,6144,0.012894222305880653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,768,0.006149333384301927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,32,0.003363555504216088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,51200,0.2565084563361274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,6144,0.03635466761059231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,10240,0.011250666446155973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,1024,0.009047110875447592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,32,0.016414221790101793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,2560,0.015083556373914083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,1024,0.012332444389661154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,6144,0.2605191071828206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,768,0.00847466703918245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,6144,0.03578222129080031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,4096,0.13079910808139378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,2048,0.04552355408668518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,16384,0.05956266985999214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,6144,0.10764444536632961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,512,0.0035546666218174826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,3072,1.0445884068806965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,64,0.012147555748621622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,6144,0.03035377793841892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,2048,0.05105066630575392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,3072,0.009504888620641496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,4096,0.05601866377724541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,2048,0.01814933286772834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,256,0.005687110953860813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,10240,0.042753779225879245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,3072,0.006439111298984951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,768,0.009149333669079674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,12288,0.27839822239345974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,10240,0.013648889131016202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,32,0.0026684444811609057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,1024,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,64,0.0024666666156715816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,65536,0.7517200046115451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,51200,2.0798070695665145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,51200,0.4360968801710341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,128,0.007975999679830339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,1536,0.046520888805389404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,256,0.024199111594094172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,1536,0.12935466236538357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,4096,0.011144888897736868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,16384,0.7136089007059733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,10240,0.0672079986996121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,3072,0.011146666275130378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,512,1.0199546813964844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,128,0.004265777766704559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,1536,0.0074133334888352295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,7168,0.24357244703504774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,1024,0.04908533228768242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,5120,0.022039999564488728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,1536,0.42010222540961367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,1024,0.010270222193664974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,5120,0.13280888398488364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,5120,0.1675928831100464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,64,0.02146311104297638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,4096,0.03618844350179037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,51200,0.1349626647101508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,256,0.005478222337034013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,51200,5.082951015896268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,8192,0.018962666392326355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,5120,0.06506133079528809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,1536,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,10240,0.02418311105834113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,1024,0.01203644441233741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,4096,0.008376888930797577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,768,0.011356444822417365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,256,0.02106311089462704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,3584,0.008849778109126622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,6144,0.010104888843165504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,6144,0.011228444675604502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,3072,0.009832888841629028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,8192,0.15870132711198595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,512,0.41898311509026426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,12288,0.32790933714972603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,4096,0.02252888845072852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,2048,0.07506666580835979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,1024,0.018592000007629395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,65536,0.0715342230266995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,2048,0.011662222445011139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,768,0.012054222325483957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,256,0.003072000005178981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,2560,0.014440000057220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,3072,0.006046222315894232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,768,0.27189334233601886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,64,0.010504888991514841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,1024,0.006174222048785951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,4096,0.013537777794731988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,1536,0.027323555615213182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,4096,0.009847111172146266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,51200,2.794929716322157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,32,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,12288,0.08754755391014947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,12288,0.04567377765973409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,12288,0.03908177879121568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,32,0.0029537777105967202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,51200,1.0278826819525826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,6144,0.16251644823286268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,32,0.004602666530344221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,768,0.21499733130137125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,32,0.00591555568906996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,3072,0.02683466672897339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,6144,0.06153688828150431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,64,0.004304888761705822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,2560,0.012187555432319641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,128,0.18800000349680582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,32,0.007366221812036302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,2560,0.00851999968290329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,7168,4.471466488308377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,128,0.006623999940024481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,5120,0.008352888955010308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,5120,0.17582221825917563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,4096,0.026369776990678575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,3584,0.15126221709781223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,6144,0.08505066898134019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,6144,0.011108444796668159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,10240,0.03234044379658169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,8192,0.011858666936556498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,512,0.0032479998966058097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,128,0.02920444475279914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,4096,0.007902222375075022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,65536,0.07231822278764513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,32,0.0914728906419542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,51200,0.6347991095648872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,3072,0.008408889174461365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,7168,0.20373688803778756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,12288,0.006963555183675554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,2048,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,128,0.1446088949839274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,1024,0.02236266599761115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,6144,0.009838222629494136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,16384,0.07124000125461154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,65536,3.5843823750813804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,128,0.0052728888889153796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,7168,0.2346044381459554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,3072,0.009593778186374241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,10240,0.07219822539223565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,51200,1.7556133270263672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,1024,0.004782222211360931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,1536,0.19776533709632027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,12288,0.9982515970865885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,51200,16.089111328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,512,0.01016266644001007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,7168,0.030380444394217596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,10240,8.196952819824219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,2560,0.12315377924177383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,51200,1.4741040335761175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,2048,0.04861066738764445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,128,0.04305155409706963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,1024,0.06335110796822442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,8192,0.2596737808651394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,1536,0.006832888970772426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,3584,0.008435555630260045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,16384,1.378351953294542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,128,0.0037751110891501107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,3072,0.013295999831623502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,128,0.02347999976740943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,64,0.009178666604889764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,8192,0.028731554746627808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,1024,0.005783111270931032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,4096,0.10891022284825642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,256,0.031854222218195595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,4096,0.006577777779764599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,8192,0.46992354922824436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,65536,0.11302399635314941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,2048,0.00981777740849389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,4096,0.023131555981106226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,8192,0.06465066803826226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,1536,0.04352533486154345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,65536,0.07827733622656928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,7168,0.010877333581447601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,1536,0.006288888967699475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,8192,0.0203128887547387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,4096,0.4688382148742676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,3072,0.0094035557574696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,32,0.015142222245534262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,512,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,2560,0.007152888509962294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,32,0.02711822258101569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,6144,0.017450667089886136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,10240,0.031798223654429116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,1024,0.15623733732435438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,1536,0.06858133607440524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,65536,1.8849474589029949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,768,0.006969778074158563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,51200,0.29944711261325413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,1024,0.006361777583758037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,256,0.0034106667670938703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,5120,0.024313777685165405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,64,0.006831999868154526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,64,0.02221511138810052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,128,0.013791999883121915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,65536,0.46917157702975804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,5120,0.009723555710580613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,768,0.018512000640233357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,8192,0.021232000655598108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,4096,0.11761511696709527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,2048,0.03701600101259019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,1536,0.004399999976158142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,10240,0.01796444422668881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,3072,0.05867377916971842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,1024,0.06489955716662936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,3072,0.3731039894951715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,256,0.003903111235962974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,2048,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,768,0.006029333505365584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,512,0.055695997344123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,12288,0.013210666676362356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,32,0.00309599998096625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,8192,0.016955556141005624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,12288,0.2710808912913005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,2560,0.11366221639845107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,7168,0.35297155380249023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,10240,0.0328800015979343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,512,0.005495111147562663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,1536,0.008662222160233391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,256,0.013807111316257052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,12288,5.2237468295627165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,64,0.010403555300500657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,32,0.014533332652515836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,7168,0.18560266494750977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,768,0.004281777888536453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,65536,0.0765191117922465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,2560,0.305165343814426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,5120,0.06966133250130548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,7168,0.0624542236328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,2560,0.0075288886825243635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,4096,0.07800977759891085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,1024,0.04327288932270474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,7168,0.02233422299226125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,256,0.005579555614127054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,2048,0.0494391123453776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,128,0.004321777572234471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,12288,0.02275644408331977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,128,0.0035297779573334586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,1024,0.016913778252071805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,3072,0.01499555508295695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,65536,0.029657777812745836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,10240,0.16820622815026176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,64,0.0035377778112888336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,3072,0.008426666259765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,51200,7.266923692491319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,7168,0.026464000344276428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,4096,0.008290666672918532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,2560,0.007704888780911763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,128,0.0063288890653186375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,10240,0.070378667778439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,1536,0.016332444217469957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,16384,1.3471999698215062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,2560,0.12116800414191352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,512,0.009518221848540837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,7168,0.013907555076811047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,256,1.8652498457166884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,1024,0.03203466534614563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,65536,3.4443653954399953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,3072,0.019268444842762418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,10240,0.007295110987292395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,256,0.00609777785009808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,5120,2.2022213406032987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,4096,0.0653217766020033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,64,0.0609804458088345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,6144,0.014202665951516895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,4096,0.019484443797005545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,2560,0.01036088913679123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,16384,0.031023999055226643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,256,0.1862515608469645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,512,0.06643378072314791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,7168,0.10110933250851101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,65536,1.563394652472602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,12288,0.08615999751620823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,10240,0.3278106583489312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,2048,0.009233777721722921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,256,0.005623999983072281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,1536,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,1024,0.014504889647165934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,10240,0.007286222444640265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,6144,0.012795555922720166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,3584,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,7168,0.045553776952955455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,65536,66.6251220703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,32,0.01647911138004727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,2560,0.1551484399371677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,4096,0.021524444222450256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,2560,0.4521244366963704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,256,0.008223999705579545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,256,0.008166222108734978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,4096,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,3072,0.009921777579519484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,51200,0.9127599928114148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,5120,0.021683555510309007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,7168,0.01647822227742937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,8192,0.3971449004279242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,768,0.04703911145528158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,6144,0.037718223200904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,1536,0.022269333402315777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,256,0.005119110975000593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,7168,1.7748151355319555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,2560,0.05343022280269199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,256,0.010183999935785929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,8192,0.16109955310821533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,2048,0.025504888759719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,51200,0.0613831082979838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,12288,0.03200444579124451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,6144,0.05608533488379585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,8192,0.5207609070671929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,32,0.0057591112951437635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,1024,0.006148444281684027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,64,0.005320888840489917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,4096,0.012489777472284106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,1024,0.8460879855685763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,12288,1.2617466184828017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,5120,0.033005333609051175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,1024,0.04831022355291578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,512,0.0037928890022966596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,65536,4.005687289767795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,7168,0.016467556357383728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,7168,0.11718311574723984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,4096,0.007315555380450354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,64,0.003546666767862108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,12288,0.2597324318355984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,10240,0.021165332860416834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,1024,0.008447999755541483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,65536,0.1068097750345866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,3072,0.01111644423670239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,3072,0.04527822136878967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,128,0.0053226666318045715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,10240,0.03926666577657064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,3584,0.03273066547181871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,3072,0.0539191100332472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,12288,0.7891715367635092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,512,0.07217866844601102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,1024,0.03747111227777269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,16384,0.1909182204140557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,1536,0.009645333720578088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,512,0.007729777859316931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,16384,0.08717688586976792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,128,0.005912888795137405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,6144,0.3367626667022705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,2048,0.007050666544172499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,1024,0.06508088774151273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,768,0.01723377737734053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,6144,0.04661866691377428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,5120,0.029016888803905908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,32,0.006268444574541516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,2560,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,65536,0.08952799770567153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,64,0.1041333344247606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,2560,0.006320000108745363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,5120,0.16808533668518066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,512,0.00683111117945777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,256,0.014865777558750577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,5120,0.011548444628715515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,8192,0.030273778571022883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,65536,1.7637351353963215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,512,0.01148799972401725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,1536,0.055141333076688975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,12288,0.007669332954618666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,51200,0.7855244212680392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,10240,0.0221413340833452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,12288,13.562295701768663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,32,0.006540444576077991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,32,0.011382222175598145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,1536,0.01073066641887029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,512,0.004846222284767362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,8192,6.955773247612847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,16384,0.6403973367479112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,2560,0.036687999963760376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,6144,0.07189422183566622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,512,0.005054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,65536,0.29878399107191295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,16384,0.023759111762046814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,12288,0.04660977919896444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,7168,0.011687999798191918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,128,0.03717422154214647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,512,0.0773093303044637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,3584,0.03451733456717597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,16384,0.8867368698120117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,128,0.006239111224810283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,3072,0.37095199690924746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,2560,1.5176702075534398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,5120,0.028100444210900202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,128,0.007115555306275685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,12288,0.07908444272147284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,1536,0.006060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,32,0.008218666745556725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,32,0.0028799999919202593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,2048,0.1340008841620551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,5120,0.01811822255452474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,1536,0.06843555635876127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,768,0.029464887248145208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,10240,0.1632924477259318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,5120,0.046112888389163546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,51200,0.2686213387383355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,6144,0.019366222951147292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,4096,0.007993777592976889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,512,0.01887999971707662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,2048,0.01648355523745219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,768,0.018560000591807894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,768,0.01461777753300137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,3072,0.009002666506502364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,128,0.026741334133678015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,1024,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,64,0.0026142222599850762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,51200,0.16016356150309244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,65536,0.06300889121161567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,8192,0.09660889042748345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,3072,0.14021421803368464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,1024,0.007892444729804993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,2560,0.009928888744778102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,5120,0.018065777089860704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,512,0.016173332929611206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,65536,0.10394133461846246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,512,0.03449689017401801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,7168,0.022407111194398668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,3584,0.007802666889296637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,32,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,16384,0.05142577820354038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,6144,0.09591733084784614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,512,0.007038222418891058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,32,0.08507821957270305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,2048,0.048531555467181735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,6144,0.019511111908488803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,2560,0.007304000357786815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,3584,0.0073520003093613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,768,0.0038364442686239877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,12288,0.007034666836261749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,3584,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,768,0.030728889836205378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,256,0.03943022092183431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,7168,1.824865765041775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,10240,0.015511110424995422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,3584,0.00793422261873881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,128,0.9301795959472656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,8192,0.016149333781666227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,256,0.0033520000676314035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,4096,0.13555110825432673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,768,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,2560,0.008759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,3072,0.014816888504558139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,3584,0.011014222270912595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,512,0.01977511081430647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,51200,2.659310234917535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,12288,0.04145688811937968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,128,0.0035573335157500375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,4096,0.016006223029560514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,4096,1.9262871212429469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,256,0.005372444374693765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,128,0.0034133332471052804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,16384,0.20101422733730742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,3584,0.1550017727745904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,16384,0.04392533169852363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,1536,0.005342222336265776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,8192,0.016198222835858662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,1536,0.01293511109219657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,5120,0.029763556189007227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,12288,0.03414577907986111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,128,0.00518488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,3072,0.130566226111518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,16384,0.030534221066368952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,10240,0.017535999417304993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,7168,0.029519110918045044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,3072,0.029679109652837116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,65536,0.6154115464952257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,8192,0.11379555861155193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,4096,0.008814222282833522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,3584,1.5620862113104925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,256,0.02943555514017741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,7168,0.012885332935386233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,7168,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,3584,0.01054666688044866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,16384,1.9533600277370875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,768,0.029899554120169744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,768,0.00810222244924969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,128,0.029554665088653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,65536,0.17952711052364775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,5120,0.08339733547634548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,4096,0.1713333394792345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,3072,0.01852799952030182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,64,0.01404266721672482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,51200,0.2770017782847087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,256,0.0053093334039052325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,65536,0.21407911512586805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,3584,0.34619911511739093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,3584,0.1051751110288832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,2560,0.01570311188697815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,3072,0.012674666941165924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,128,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,3072,0.0074577778577804565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,12288,0.043896890348858304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,512,0.14642043908437094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,64,0.0028115556471877625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,64,0.002806222273243798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,768,0.08559911118613349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,7168,0.02316444449954563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,10240,0.07507999738057454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,3584,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,256,0.10453689098358154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,1024,0.036122666464911565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,1536,0.015698666373888653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,512,0.01752800080511305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,128,0.007653333246707916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,2048,0.10304711262385051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,8192,0.1570684380001492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,5120,0.07265688975652058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,1536,0.20155910650889078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,512,0.03988088832961188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,512,0.005699555493063397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,2048,0.04654755526118808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,51200,0.13174488809373644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,16384,1.3193448384602864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,3072,0.20092266135745576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,65536,0.025292444560262892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,3584,0.031904889477623835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,128,0.008914666871229807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,10240,0.030889779329299927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,10240,0.08518489201863606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,12288,0.021939555803934734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,10240,0.01384622189733717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,4096,0.7222222222222222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,3584,0.04325422313478258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,2560,0.00869333330127928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,3072,0.12536711162990993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,5120,0.5666880077785915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,1024,0.008751110898123847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,12288,0.5241164631313747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,3072,0.5428275532192653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,1024,0.009635555247465769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,7168,0.45944976806640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,2048,0.00815644446346495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,65536,0.5093208948771158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,65536,0.34527288542853457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,3584,0.018546667363908555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,1536,0.38872088326348203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,128,0.0029395555870400537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,3072,0.007069333559936947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,128,0.009343110852771336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,12288,0.010042666561073726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,16384,0.01072266697883606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,256,0.005940444353553984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,1536,0.006808888994985157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,4096,0.03047377864519755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,51200,0.10347911384370591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,12288,0.04888088835610283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,12288,0.012749333348539142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,6144,0.013632000320487551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,65536,0.8643057611253527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,3072,0.1809093289905124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,768,0.011523555550310345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,51200,0.014659555421935188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,5120,0.015257777439223396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,128,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,3584,0.00795733349190818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,16384,0.07304088936911689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,512,0.026039999392297532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,1536,0.01017777787314521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,256,0.016344888342751395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,1536,0.046367999580171376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,2048,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,1024,0.006383111079533895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,1536,0.006890666567616993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,1024,0.07668444183137682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,3072,0.581759982638889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,3072,0.023710222707854375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,12288,0.39142844412061906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,5120,0.07428355349434747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,12288,0.021199999584092036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,2560,0.013157332936922709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,10240,0.012141333685980903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,4096,0.09324000279108684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,512,0.005981333139869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,3584,0.15534310870700413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,3072,0.1797626680798001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,51200,0.5476177533467611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,1536,0.03728888763321771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,64,0.0029120000286234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,6144,0.01574488812022739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,51200,2.027893278333876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,1024,0.005426666802830166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,32,0.012879110872745514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,512,0.010596444209416708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,1536,0.024951110283533733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,65536,45.47520955403646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,1536,0.008172444171375698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,768,0.006490666833188798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,10240,0.27049244774712455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,65536,0.1893493334452311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,256,0.09896088971032037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,4096,0.009029332962301042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,128,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,3584,0.012272888587580787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,1024,0.010205333431561789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,128,0.013815110756291283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,6144,0.013016888664828407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,6144,0.22284356753031412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,6144,0.03781333234575059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,51200,0.196070220735338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,512,0.005680000202523337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,32,0.0028648889727062653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,3584,0.007848000360859765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,128,0.008599110775523717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,256,0.002875555513633622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,512,0.04709777898258633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,32,0.014369777507252164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,3072,0.006212444355090459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,256,0.023338667220539514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,64,0.015248000621795654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,2048,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,4096,0.1330666674507989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,3584,0.14011200269063315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,5120,0.13892000251346165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,2048,0.006890666567616993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,2560,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,32,0.02633511192268795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,7168,0.2435528967115614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,6144,0.016209777858522203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,16384,0.02529688841766781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,128,0.005241777747869492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,3072,0.02142400046189626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,2560,0.06848888927035861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,512,0.00554933316177792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,65536,0.2730044523874919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,16384,0.07925688558154635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,1536,0.013981334037250943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,10240,0.1630800035264757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,7168,0.025794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,1024,0.026001777913835313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,8192,0.030208888981077407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,12288,0.045702222320768565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,2560,0.009791110952695211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,512,0.01997244523631202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,64,0.027445332871543035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,2048,0.32183554437425405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,768,0.011257777611414591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,3072,0.14638488822513157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,51200,0.13725599977705213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,2048,0.02201244400607215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,16384,0.024933333198229473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,32,0.005879111174080107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,64,0.002623999905255106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,4096,0.020297777321603563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,16384,12.611663818359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,3072,0.0706604454252455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,128,0.004129777765936322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,1536,0.011183111204041375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,256,0.8158577813042535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,65536,0.5871244536505805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,51200,11.519593132866753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,5120,0.13759911060333252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,128,0.05579821930991279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,2560,0.0744311081038581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,64,0.002549333290921317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,12288,2.023415035671658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,5120,0.016599110431141324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,256,0.03730577892727322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,2048,0.01927288870016734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,3584,0.022287999590237934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,3584,0.01105955574247572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,4096,0.06790666447745429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,8192,0.030935111973020766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,10240,0.6763555738661023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,512,0.007999111380841997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,1536,0.011065777805116443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,1024,0.011071999867757162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,16384,0.02996622191535102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,16384,0.13267910480499268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,6144,0.16606933540768093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,7168,0.019398222366968792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,128,0.0033866665843460294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,2048,0.16500000158945718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,256,0.19399821758270264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,12288,0.07070133421156141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,256,0.0047528888616296975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,768,0.026334222820070054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,10240,0.08024977975421481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,512,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,5120,0.0063662222690052455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,7168,0.021724444296624925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,16384,0.04898400108019511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,2560,0.008039111064540016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,7168,0.020761777957280476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,256,0.004785777793990241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,16384,0.06195200151867337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,6144,0.5473359955681695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,768,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,4096,0.290720886654324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,10240,0.5478906631469727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,768,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,256,0.004135999828577042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,16384,0.4101848867204454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,12288,0.6986142264472114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,5120,0.04324533210860359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,65536,1.9595955742730036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,4096,0.014780445231331719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,128,0.03142044279310439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,1024,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,5120,0.013459555804729462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,10240,0.10710489087634617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,4096,0.30860532654656303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,32,0.0028728888266616394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,4096,0.051085333029429116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,1536,0.1598773267534044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,12288,0.017690666847758822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,2560,0.006741333338949416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,128,0.007479999628331926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,6144,0.021270222134060327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,2560,0.006093333164850871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,2560,0.13559377193450928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,65536,0.01386844449573093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,5120,0.009698666632175446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,10240,0.7274275355868869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,1024,0.1807911131117079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,65536,0.2573022312588162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,512,0.007385777930418651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,51200,0.13605333699120417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,2048,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,768,0.023899555206298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,1536,0.009071110851234859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,51200,0.06219644678963555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,4096,0.01608800060219235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,512,0.024706666668256123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,2560,0.009415999882751042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,10240,0.24604800012376574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,1024,0.01940711173746321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,51200,0.09599110815260146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,512,0.011622221933470832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,2560,0.00981600003110038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,3584,0.13150310516357422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,8192,0.005879111174080107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,8192,0.015232889188660515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,1536,0.035543110635545515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,768,0.979948467678494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,1024,0.020844444632530212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,768,0.018724444839689467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,6144,0.05815911293029785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,2048,0.01051999959680769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,2560,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,4096,0.008809777597586313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,512,0.02110933264096578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,6144,0.43374310599433047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,1024,0.013241777817408243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,10240,0.027057778504159715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,64,0.002598222138153182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,32,0.0026702222724755607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,768,0.005848888721730974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,4096,0.020579554968410067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,3584,0.09305333428912693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,4096,0.026012443833880957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,3584,0.06493422057893541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,65536,0.6704826884799533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,16384,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,12288,0.010235555469989777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,2048,0.016599999533759225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,6144,0.0057884446448749965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,2560,0.06568977567884657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,32,0.005183111048407025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,6144,0.01684177749686771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,6144,0.00794488853878445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,5120,0.008578666382365758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,2048,0.10886577765146892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,2048,0.08469066354963516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,4096,0.04452711012628344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,12288,0.39251822895473903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,128,0.013349333571063148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,10240,0.017679111825095285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,5120,0.02769688930776384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,6144,0.018152889278199937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,2560,0.006791111081838608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,7168,0.02102488941616482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,16384,0.021094222863515217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,65536,0.15528178215026855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,3584,0.0079644446571668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,512,0.014666666587193808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,8192,4.592414432101779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,16384,0.1690044403076172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,3072,0.011693333586057028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,7168,0.03849600089920892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,768,0.005605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,128,0.007800888684060838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,1536,0.05739911397298177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,3072,0.013930666777822705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,16384,0.07473511166042753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,512,0.013110222087966071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,12288,0.2777715524037679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,6144,0.01035733355416192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,64,0.013607111242082385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,2048,0.028607110182444256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,4096,0.012537777423858643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,3072,0.00926577796538671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,1024,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,2560,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,2048,0.034259554412629865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,2048,0.03753066725201077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,7168,0.189428448677063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,10240,1.8144444359673395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,1536,0.01218488895230823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,512,0.003678222083383136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,16384,0.006594666590293248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,7168,0.15257244639926487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,3072,0.0796497795316908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,7168,0.8772666719224719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,256,0.007734222544564142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,3072,0.04443822304407755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,7168,0.0863422221607632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,10240,0.01574133336544037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,256,0.027121777335802715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,768,0.01945333348380195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,2560,0.2571013238694933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,128,0.020896888441509668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,51200,0.17515199714236787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,4096,0.020192000601026747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,512,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,5120,0.21207288901011148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,3584,0.18247644106547037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,10240,0.16531911161210802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,512,0.004765333400832282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,32,0.005094222310516569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,51200,0.3763679928249783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,3072,0.14211555322011313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,51200,72.59598795572917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,32,0.0031804444475306403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,3072,0.05200799968507555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,1536,0.29352090093824595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,51200,14.224917941623263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,128,0.007270221908887227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,128,0.03056266572740343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,16384,0.252799113591512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,8192,0.01815022197034624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,16384,0.20236978265974256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,512,0.014141332772043018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,512,0.005064888960785336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,5120,0.011703111231327057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,3584,0.06982577509350248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,64,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,256,0.028138667345046997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,256,0.12203644381629096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,2560,0.055161777469846934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,6144,0.3699768914116754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,8192,0.0169804443915685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,16384,0.07593688699934217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,12288,0.2586053212483724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,16384,0.034727109803093806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,1536,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,51200,0.08514577812618679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,512,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,6144,0.2572568787468804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,10240,0.5975848833719889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,6144,0.014518222875065274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,12288,0.3954462210337321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,10240,0.035954667462242976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,3584,0.34012354744805234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,128,0.003050666716363695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,16384,0.02198933396074507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,512,0.0036764442920684814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,128,0.003395555540919304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,5120,0.043974220752716064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,3072,0.04239644606908163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,256,0.025242666403452556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,12288,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,5120,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,2560,0.10575377941131592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,4096,0.061194664902157254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,1024,0.07565510935253568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,2048,0.006551110910044775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,1024,0.03268622358640035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,128,0.015046222342385186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,16384,0.036496000157462224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,16384,1.3909218046400282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,5120,0.10619466834598118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,5120,0.06756444772084554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,10240,0.021407999926143225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,3584,0.03380089004834493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,12288,0.20163733429378936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,2560,0.47286754184299046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,10240,0.05993777513504028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,1024,0.01793866687350803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,5120,0.012919999659061432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,128,0.003170666595300039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,8192,0.03153422143724229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,6144,0.05703910854127672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,51200,0.015391111373901367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,4096,0.0248026665714052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,512,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,6144,0.010126222338941362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,12288,0.05820177661048042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,128,0.0063235556913746735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,2560,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,128,0.006229333165619109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,2560,0.0076453329788313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,512,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,3072,0.051206221183141075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,3072,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,1024,0.004676444249020683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,128,0.13983555634816489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,512,0.003104888937539525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,2048,0.014388443695174323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,10240,2.16374757554796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,3072,0.016539555456903245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,2560,0.03651911020278931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,1536,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,512,0.007903999752468532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,32,0.0036106668412685394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,3584,0.16635999414655897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,8192,0.06697599755393134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,128,0.004238222208287981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,64,0.007431999676757389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,256,0.03787644373046027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,64,0.00554577757914861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,768,0.04065510961744521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,32,0.0028968888024489083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,1024,0.11632800102233887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,6144,0.2775546709696452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,3072,0.19519555568695068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,4096,0.007040888898902469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,65536,3.229108386569553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,64,0.0030497778207063675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,51200,0.06365511152479383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,2560,0.08783911334143744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,1536,0.013017777767446307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,65536,0.09624355369144016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,7168,0.013853333062595792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,768,0.004015111260943943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,5120,0.011992000043392181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,3072,0.010649777948856354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,8192,0.014567111929257711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,6144,0.06351288821962145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,65536,0.5790808995564779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,1024,0.012818666795889536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,512,0.009831111464235518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,6144,0.00998933364947637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,6144,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,128,0.005285333428117964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,2560,0.00703288863102595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,1024,0.019288889235920377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,4096,0.019018666611777413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,768,0.022227555513381958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,6144,0.06334755818049113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,12288,0.02404977712366316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,8192,0.2655777666303846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,64,0.00535733335547977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,12288,0.48557154337565106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,1024,0.014597333139843412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,8192,0.013040000365840064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,8192,0.044934223095575966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,65536,2.716920852661133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,7168,0.027531555957264368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,1536,0.00886577781703737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,2560,0.025124443901909724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,1536,0.0074515557951397365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,3584,0.007094221810499827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,1536,0.030163556337356567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,4096,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,128,0.005031110925806893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,12288,0.010416888528399997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,1024,0.0224906669722663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,5120,0.400273773405287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,256,0.0034044444974925783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,512,0.017130666308932833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,5120,0.02459555533197191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,1536,0.0852960016992357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,12288,0.028218666712443035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,1024,0.09582310914993286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,5120,0.21138311756981742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,7168,0.6029626528422037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,6144,0.016851555969980028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,7168,0.19726222091250947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,16384,0.022297778063350256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,256,0.011260444091426002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,51200,3.67142825656467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,1024,0.02956177790959676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,256,0.0047102222839991255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,2048,0.020999110407299463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,51200,1.6603227191501195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,6144,0.034860445393456355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,16384,0.026736888620588515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,65536,3.9798677232530384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,65536,0.11049955421023899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,12288,0.1363688839806451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,4096,0.7548311021592883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,128,0.016325333052211337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,512,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,65536,0.8377048704359266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,128,0.013643555343151093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,64,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,32,0.00794488853878445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,128,0.0041368889311949415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,65536,1.2206978268093533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,4096,0.022315555148654517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,65536,0.14386399586995444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,1536,0.008047999607192146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,65536,0.560381359524197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,6144,0.050373332368003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,768,0.010675555302037133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,128,0.009574222067991892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,3072,0.014155555102560254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,1024,0.00620266670982043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,768,0.004328000048796336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,4096,0.03154933452606201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,128,0.004921777794758479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,7168,1.2486613591512044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,64,0.009540444446934594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,1536,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,1024,0.010252444280518426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,8192,0.2134711080127292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,128,0.00351555562681622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,256,0.005240889059172736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,4096,0.4650017950269911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,256,0.013239999612172445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,6144,0.006441777778996362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,7168,0.1475528875986735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,3584,4.933953603108724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,51200,0.13283467292785645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,6144,0.2357093228234185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,768,0.006893333461549547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,51200,0.06425066788991292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,12288,0.052024000220828585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,256,0.003350222276316749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,1024,0.01520088811715444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,16384,0.07604089048173693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,768,0.021708443760871887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,1024,0.015860444969601102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,6144,0.05003644360436333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,5120,0.011787555283970304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,2048,0.010771555205186209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,1536,0.006503110958470239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,3072,0.04714666803677877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,8192,1.5136249330308702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,3584,0.03170399864514669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,3072,0.01919555498494042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,2560,0.03767555620935228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,5120,0.13916533523135716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,2560,1.7358267042371962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,6144,5.3086730109320746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,7168,0.03567466802067227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,512,0.01092977739042706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,7168,0.22179110844930014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,3584,0.008512000242869059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,16384,0.529927995469835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,768,0.008231999973456064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,51200,0.4542186525132921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,65536,5.034717135959201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,10240,2.526543935139974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,2560,0.009535111486911774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,1024,0.016971555021074083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,1024,0.006968888971540663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,32,0.0027884443600972495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,65536,0.12460622522566055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,3584,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,6144,0.026657778355810378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,51200,0.864866680569119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,256,0.014679110712475248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,3584,0.10643021927939521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,6144,0.01683999929163191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,512,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,1536,0.24985154469807944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,128,0.015444444285498725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,256,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,3072,0.1323688824971517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,7168,0.024705777565638225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,10240,0.18381778399149576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,1024,0.009519110951158736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,2048,0.052984890010621816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,1536,0.009706666900051964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,8192,0.021391110287772283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,6144,0.022506665852334764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,768,0.005480000128348668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,7168,0.040376000934176974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,64,0.0027466666781240036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,65536,0.6036808755662706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,3584,0.009195555415418413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,2048,0.020023999942673575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,32,0.01148977792925305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,51200,0.1112053394317627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,51200,0.06031555599636502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,4096,0.029438220792346533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,2560,0.08181866672303942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,2560,0.00611200018061532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,16384,0.027679110566775005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,51200,0.6053146786159939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,16384,5.261175791422526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,16384,0.027131555808915034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,512,0.005079111291302575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,128,0.0034515555534097883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,65536,2.0911839803059897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,512,0.044616887966791786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,32,0.002622222320901023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,1536,0.004818666726350784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,16384,0.05636622508366903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,8192,0.42251110076904297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,10240,0.05156177613470289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,8192,0.01443911095460256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,3584,0.11359555191463894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,10240,0.006666666517655055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,7168,0.016398222910033334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,3584,1.0541760126749675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,4096,0.036213331752353244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,3072,0.006450666735569636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,10240,0.4988035625881619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,10240,0.013565333353148567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,6144,0.4454951286315918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,3584,0.009407999614874521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,8192,0.015972443752818637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,1024,0.0347360008292728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,10240,0.03571644425392151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,10240,0.19928977224561903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,3072,0.01275022245115704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,65536,0.11603200435638428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,128,0.004906666775544484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,2560,0.21172532770368788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,3072,0.04633511106173197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,12288,0.09426399734285142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,64,0.004825777891609403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,12288,0.1555697785483466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,65536,0.9103920194837781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,1024,0.005629333357016246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,3072,0.0105608892109659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,10240,0.07396088706122504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,65536,0.009849777652157677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,1536,0.02464977734618717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,4096,0.01606044504377577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,64,0.002604444407754474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,3584,0.007856888903511895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,51200,0.28073867162068683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,1024,0.09866488642162746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,768,0.1082693338394165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,2048,0.009099555512269339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,51200,0.06226133637958103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,32,0.14656355645921496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,12288,0.01819644371668498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,12288,0.031317333380381264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,32,0.008808888494968414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,128,0.002728888971938027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,12288,0.30115199089050293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,768,0.004619555754794015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,16384,5.663488176133897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,4096,0.01418133411142561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,7168,0.06496444675657484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,1024,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,128,0.004405333350102107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,1536,0.0075911109646161394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,1024,0.045548445648617215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,3584,0.16690666145748564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,3072,0.17145778073204887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,3072,0.09559644593132867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,65536,2.0516346825493708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,6144,0.043219556411107383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,1024,0.04384444488419426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,256,0.03103733393881056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,65536,0.015772443678643968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,2560,0.00793866647614373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,256,0.00591111100382275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,16384,0.28233154614766437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,768,0.0034924443397257063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,12288,2.813448799981011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,3584,0.014304889572991265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,512,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,10240,0.03074399961365594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,256,0.0063902222447925145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,5120,0.02455377744303809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,1024,0.03444000085194906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,3072,0.018041777941915724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,2560,0.05936888853708903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,1536,0.012157333393891653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,512,0.007242666350470648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,16384,0.020383111304706998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,3584,0.014007111390431723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,65536,0.07218844360775419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,7168,0.01201600001917945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,768,0.008382221890820397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,65536,0.08420444197124904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,128,0.007041778001520369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,3072,0.0120328888297081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,2048,0.08915999862882827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,65536,0.06342755423651801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,2560,0.007151111132568783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,2560,0.021341333786646526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,512,0.023381332556406658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,6144,0.07031110922495525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,3072,0.00775822252035141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,8192,0.027898665931489732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,2048,0.07784444755978055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,12288,0.040017777019076876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,7168,0.023005333211686876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,16384,0.05996977620654636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,12288,0.030660443835788306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,8192,0.05334844523006015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,512,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,1024,0.013740444348918067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,4096,0.5473875469631618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,3584,0.017468444175190397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,65536,0.1861422194374932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,5120,0.02384888960255517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,5120,0.01184977756606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,3584,0.013764444324705334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,1536,0.0660346680217319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,7168,0.2441502147250705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,1024,0.0251884443892373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,128,0.004502222355869082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,512,0.00572622236278322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,128,0.008396444221337637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,1536,0.013180444637934366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,2048,0.00998755544424057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,128,0.005169777820507686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,16384,1.082765367296007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,16384,0.023312888211674158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,51200,0.14773333072662354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,51200,1.2396684222751193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,12288,0.09138755665885077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,2048,0.0064497776329517365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,256,0.003083555648724238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,768,0.0878702203432719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,7168,0.01924977699915568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,256,0.007788444558779399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,3584,0.01423555612564087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,6144,0.010967111421955956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,1024,0.003698666476541095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,256,0.026092444856961567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,10240,0.1991057793299357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,128,0.003847111016511917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,4096,0.009476444787449306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,8192,0.4931031333075629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,65536,2.7075697580973306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,2048,0.020226667324701946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,5120,0.03391999999682108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,2048,0.09077688720491196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,64,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,8192,0.013569778038395775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,1536,0.006821333534187741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,16384,0.05240799983342489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,5120,0.022805333137512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,51200,0.10148444440629746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,4096,0.008309333688682979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,4096,0.06732800271775988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,768,0.1922008858786689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,2048,0.005544888890451855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,4096,0.00813155538505978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,4096,0.19030043813917372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,2560,0.01257066677014033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,16384,0.14552888605329725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,2048,0.017588444881969027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,65536,0.009841777384281158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,16384,0.4397093454996745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,65536,0.12515466743045384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,768,0.008728889127572378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,7168,0.012614222036467658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,1536,0.008296888735559251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,16384,1.2128719753689237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,3584,0.04119111100832621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,10240,0.03191555539766947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,512,0.008230222596062554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,3072,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,65536,0.40346667501661515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,256,0.006471111128727595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,5120,0.008062221937709385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,8192,0.019066666563351948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,1536,0.07197511196136475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,2048,0.10940000083711411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,3584,0.013450666434235044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,7168,0.028199109766218398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,7168,0.015551111764378019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,6144,0.432328012254503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,3584,0.0081395556529363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,4096,0.01609777741962009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,128,0.005101333475775189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,16384,0.279745790693495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,512,0.024340444140964087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,16384,0.03993599944644504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,2048,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,512,0.004620444443490771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,2560,0.042488889561759106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,512,0.005944889038801193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,256,0.04491288794411553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,3072,0.6994435522291395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,3072,0.02427022159099579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,2048,0.05950578053792318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,1024,0.007989333735571967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,10240,0.13421600394778782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,7168,0.030864887767367895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,1536,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,32,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,1024,0.04062044289377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,4096,0.009183111290136972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,512,0.02347466680738661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,1536,0.05172266562779745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,1024,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,51200,1.4977315266927083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,3072,0.03053511182467143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,256,0.006864000111818314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,4096,0.04789688852098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,128,0.0029440000653266907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,51200,0.06464711162779066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,6144,0.011454222102959951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,1536,0.059471997949812144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,51200,4.480998145209418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,1536,0.04885866575770908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,6144,0.012708444562223224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,3072,0.013387555877367655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,3584,0.007544889218277401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,32,0.05917955769432915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,64,0.16705421606699625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,128,0.007038222418891058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,4096,0.054882665475209556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,768,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,12288,0.056384887960222035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,1024,0.25309422281053334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,2048,0.0070293330483966405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,51200,0.20425332917107475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,5120,2.591545740763346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,7168,2.1997670067681208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,1024,0.1302622159322103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,3072,0.1237306727303399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,768,0.03884622123506334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,7168,0.016453334026866488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,1024,0.025270221961869135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,8192,0.032991111278533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,256,0.009896000226338705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,5120,0.2712542215983073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,4096,0.008967999782827165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,2560,1.0401333702935112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,1024,0.008226667013433244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,8192,0.020744888318909537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,12288,0.027870221270455256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,51200,2.3010391659206815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,65536,0.13592888249291316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,32,0.004384888956944148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,5120,0.16305511527591282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,5120,0.032688889238569475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,3072,0.15776799784766302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,32,0.01460355520248413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,64,0.0031724443866146933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,128,0.005698666804366642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,3584,0.009229333036475712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,128,0.014520888527234396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,3584,0.12809422281053331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,2048,0.019855111837387085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,128,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,256,0.004384888956944148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,128,0.007455999652544658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,2048,0.011046222514576383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,3584,0.006726222319735422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,10240,0.014337778091430664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,7168,0.03244533472590976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,16384,3.154905743069119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,2560,0.12666311528947619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,10240,0.029366221692827012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,8192,0.039148443275027804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,2560,0.006671999891599019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,5120,0.021847999758190576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,3072,0.015031110909250049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,5120,0.1676559978061252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,12288,0.03449689017401801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,3584,0.025695110360781353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,256,0.0029271110478374693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,256,0.004599110947714912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,1536,0.09098488754696316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,7168,0.5406995349460179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,2048,0.014543111125628153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,10240,0.019341333044899833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,2048,0.02857066525353326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,1024,0.020478222105238173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,512,0.006433777925040986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,1536,0.04260622130499946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,10240,0.015277332729763456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,16384,4.190565321180555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,3072,0.017078222499953378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,32,0.01131377783086565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,768,0.08543199963039821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,51200,0.3061537742614746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,51200,0.519719123840332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,10240,0.16339022583431667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,10240,0.009962666365835402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,7168,0.012792000340090858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,8192,0.15138667159610325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,5120,0.027847111225128174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,1536,0.009196444518036312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,8192,0.014616888430383472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,64,0.004839111119508743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,3072,0.8138160175747342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,768,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,65536,0.026963555150561865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,2560,0.2829502158694797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,1536,0.016162667009565566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,16384,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,6144,0.020119110743204754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,16384,0.23760178354051378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,8192,0.3225875430636936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,1024,0.011295999917719098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,65536,0.15819556183285183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,1536,0.020689777202076383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,1024,0.008823999928103553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,2048,0.1122515598932902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,1024,0.01274311128589842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,512,0.004022222426202563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,1024,0.00923733330435223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,512,0.012135999898115793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,8192,0.01828799976242913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,10240,0.020456888609462313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,6144,0.04811199837260776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,10240,0.04974755644798279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,16384,0.03004266818364461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,5120,0.01696711116366916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,1024,0.012795555922720166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,16384,0.27756532033284503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,3072,0.009217778013812171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,65536,2.3682488335503473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,2560,0.06316177712546454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,12288,0.5232693354288737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,65536,2.8591662512885203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,12288,0.04218488931655884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,768,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,128,0.004954666727119022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,6144,0.017673777209387887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,4096,0.04605866803063286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,8192,0.40486934449937606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,4096,0.00908622228437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,768,0.02267733381854163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,1024,0.5456586413913304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,8192,0.020572443803151447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,16384,0.26759645673963756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,1536,0.01734844512409634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,51200,3.5757954915364585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,768,0.09750666883256699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,1024,0.003944000022278892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,1024,0.3827253447638617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,256,0.03367644548416138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,5120,0.014241778188281588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,65536,1.4805271360609267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,12288,0.377783113055759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,3584,0.019166222876972623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,51200,0.08712000317043728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,512,0.047762665483686656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,4096,0.025097777446111042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,5120,0.29912177721659344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,1024,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,4096,2.7241528828938804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,65536,0.3706835640801324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,51200,7.999803331163194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,5120,0.01982666717635261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,51200,2.413426717122396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,5120,0.012910222013791403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,6144,0.5305351151360406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,4096,0.011776000261306763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,16384,0.10405688815646702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,128,0.008943999807039896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,16384,0.04642311069700453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,2560,0.29049332936604816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,32,0.005467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,5120,0.21632888582017687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,768,0.00360622215602133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,10240,0.16156178050571018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,65536,11.341195848253038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,4096,0.024327110913064744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,128,0.00868444475862715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,3584,0.01869066721863217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,128,0.012111110819710625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,16384,0.0388408899307251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,32,0.01198933356338077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,10240,0.19819111294216582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,5120,0.03866044349140591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,2048,0.006555555595291986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,4096,0.13869955804612902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,3584,0.007290666302045186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,512,0.0042364444169733255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,2048,0.015651555524932016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,6144,0.041314668125576444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,64,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,16384,0.1804880036248101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,4096,0.016928889685206942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,10240,0.2603306770324707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,2560,0.005715555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,5120,0.013438222308953604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,7168,0.04731733269161648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,3584,0.05801333321465386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,3072,0.007228444847795699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,32,0.0025617778301239014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,8192,1.1947413550482857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,16384,0.03020533257060581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,256,0.0035084444615576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,3072,0.022445332672860887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,5120,0.015970667203267414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,64,0.007772444850868649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,65536,0.14644977781507704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,51200,0.09663377867804633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,10240,3.5134160783555775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,10240,0.019136889113320243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,16384,6.887650807698567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,10240,0.026543999711672466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,8192,0.11786311202579075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,768,0.009679111341635386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,1536,0.031118220753139917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,3072,0.027458666099442378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,3584,0.008124444219801161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,1024,0.060807108879089355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,6144,0.013890667094124688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,3072,0.01695644524362352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,512,0.003524444583389494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,6144,0.24531732665167916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,7168,0.06779822376039293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,128,0.007152888509962294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,10240,0.04924799998601278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,8192,0.01163644426398807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,5120,0.007276444799370236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,8192,0.21001866128709581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,12288,0.014911999305089315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,6144,0.030339555607901678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,2048,0.6071928872002496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,1536,0.05927111042870415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,51200,0.20415822664896646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,256,0.005277333160241445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,65536,0.7396533224317762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,512,0.044770667950312294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,8192,0.022321777211295232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,3584,0.10406400097741021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,128,0.007445333732499017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,5120,0.02164977788925171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,32,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,8192,0.08539377980762058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,4096,0.008970666262838576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,16384,0.028336889213985864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,5120,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,6144,0.019342222147517733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,64,0.007002666592597961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,3584,0.015260444747077094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,768,0.010962666736708747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,2560,0.022124444444974262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,768,0.08719822433259751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,3072,0.3421271112230089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,10240,0.08434666527642144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,3584,0.031210667557186548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,16384,1.1983653174506295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,65536,0.08873066637251113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,65536,0.7881235546535916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,3584,0.016420443852742512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,7168,0.011680888632933298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,768,0.006595555692911148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,10240,3.4902532365587025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,768,0.012511110968059964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,10240,0.02587022218439314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,64,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,6144,5.563966115315755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,1536,0.009517333573765224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,2048,0.11601066589355469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,5120,0.15413422054714626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,65536,0.07132977909511991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,1024,0.04126400086614821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,256,0.0033644443998734155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,6144,0.012004444168673621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,3584,0.049384001228544444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,3072,0.036624001132117376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,12288,0.5963430934482151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,12288,0.12426400184631348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,10240,0.17565155029296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,7168,0.015750222735934787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,512,0.027407111393080816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,128,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,64,0.0052720002002186244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,768,0.008904889225959778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,768,0.0391537778907352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,10240,0.015585778488053216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,3584,0.01382844481203291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,1536,0.014913777510325113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,5120,0.030849777989917334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,2560,0.013456000222100152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,2048,0.009383999639087254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,2560,0.04464800159136454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,1536,0.023295111126369897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,2048,0.004962666581074397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,4096,0.02607466611597273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,16384,0.014195554786258273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,51200,13.364855448404947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,1536,0.009513777991135916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,12288,0.21058221658070883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,3584,13.898921542697481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,12288,0.05537155601713392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,256,0.006148444281684027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,7168,0.03421155611673991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,4096,0.020462221569485135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,65536,0.08164711130989923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,1536,0.015293333265516492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,65536,0.23893600040011934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,12288,0.031191110610961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,10240,0.14830311139424643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,4096,0.04910133282343546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,1024,0.007956444389290279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,64,0.005403555515739653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,512,0.005510222166776657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,7168,0.012646222280131446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,2048,0.007103111180994246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,1024,0.006981333096822103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,16384,0.241958220799764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,128,0.012423999607563019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,2048,0.0069884442620807225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,3072,0.016321778297424316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,3072,0.3345191213819716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,512,0.007479999628331926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,65536,0.022878222995334204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,768,0.0046364445653226645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,5120,0.10064178042941624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,32,0.00554044461912579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,64,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,2560,0.03686577743954129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,4096,0.060137775209214955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,768,0.0184515549076928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,10240,0.02278933260175917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,3072,0.14319645033942327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,51200,4.124905056423612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,5120,0.06320089101791382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,512,0.028799109988742407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,6144,0.013276444541083442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,5120,0.05564888980653551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,3584,0.01630577776167128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,3072,0.0127697777416971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,8192,0.4766346613566081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,3584,1.1444390614827473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,65536,0.07543200254440308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,3584,0.01511111193233066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,32,0.0052151112920708126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,7168,0.01759022143152025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,4096,0.021987555755509272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,32,0.0027173333283927706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,16384,0.03755022088686625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,32,0.003134222287270758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,65536,0.12434844175974528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,2048,0.007070221834712558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,2048,0.03473155697186788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,768,0.01979377700222863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,8192,0.0075128889746136135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,2560,0.010232888989978367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,3584,0.02700800034734938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,512,0.004429333325889376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,768,0.026886221435334947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,8192,2.8237262301974826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,3072,1.3022008472018771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,16384,0.14842132727305093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,6144,0.0413991113503774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,8192,0.11165066560109456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,4096,0.009436444275909001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,2560,0.020192889703644645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,1024,0.015557333827018738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,3584,0.0379617777135637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,7168,0.059246222178141274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,2048,0.04070666763517592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,10240,0.011698666546079846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,32,0.0025804444319672054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,51200,0.13976532883114284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,256,0.009143999881214565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,51200,2.4515048133002386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,2048,1.7750924428304036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,10240,1.0570702022976346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,1024,0.05182577835188972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,8192,0.21174666616651747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,6144,0.09689866834216648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,256,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,12288,0.016395555602179635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,4096,0.009454222189055549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,16384,6.4667875501844625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,3584,0.010739555789364709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,51200,0.06261866622500949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,2560,0.023388443721665278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,512,0.005719111197524601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,512,0.008364444805516137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,256,0.05196533269352383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,7168,0.029679109652837116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,768,0.014458666245142618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,12288,0.011229332950380115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,768,0.047713776429494224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,4096,0.14176978005303278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,3584,0.007927999728255803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,256,0.003850666599141227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,768,0.004624888714816836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,512,0.006815111057625876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,65536,0.42284178733825684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,2560,0.08308266931109957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,64,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,12288,0.3751653300391303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,1024,0.016373333003785875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,8192,0.011528000235557556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,3584,0.39101333088344997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,32,0.0035617777870761026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,7168,1.1101901796129015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,7168,0.25038401285807294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,51200,0.5358106825086806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,12288,0.022084444761276245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,3072,0.08666133218341404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,2560,0.01293511109219657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,2048,0.010584889186753167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,3072,0.0069226668112807805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,3072,0.007862221863534715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,2048,0.07643644677268134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,2048,0.14929244253370497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,6144,0.06572888957129584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,3072,0.05820710791481865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,6144,0.15537688467237684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,64,0.047726223866144814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,65536,7.941162109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,8192,5.829814910888672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,4096,0.025755556093321905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,128,0.010410666465759277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,1024,0.04031999905904134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,12288,0.2553617689344618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,51200,0.155967116355896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,7168,0.052321778403388135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,6144,0.027689778142505225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,12288,0.6081324683295356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,10240,0.40077779028150773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,16384,0.2893288930257161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,2048,0.50527556737264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,12288,0.4039155642191569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,4096,0.010300444232092964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,64,0.005388444496525659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,12288,0.19806222120920816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,768,0.011760888828171624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,8192,0.011801777614487542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,2560,0.019290667441156175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,1536,0.010570666856235929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,51200,8.044039408365885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,128,0.01650311052799225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,256,0.03639555639690823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,1024,0.03423377871513367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,8192,0.09717066420449151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,256,0.003900444342030419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,65536,0.23687820964389375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,51200,2.4330897861056857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,51200,0.07922933499018352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,5120,0.11692710717519124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,4096,0.2814284430609809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,8192,0.276530663172404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,128,0.010022222167915767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,51200,6.658731672498916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,16384,0.008738666772842407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,10240,0.04552888870239258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,2048,0.07811022467083402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,10240,0.017987555927700467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,128,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,8192,0.011926222178671094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,10240,0.026506667335828144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,51200,0.05180711216396756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,1536,0.00685511115524504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,32,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,512,0.008463111188676622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,512,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,3584,0.06451377603742811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,6144,0.09091377920574611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,128,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,6144,0.03584355447027419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,768,0.029880000485314265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,5120,0.1918568876054552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,1536,0.005544888890451855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,3072,0.07333778010474311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,7168,0.025614221890767414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,2048,0.0843377775616116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,7168,0.01350844403107961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,768,0.011113777756690979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,1024,0.005842666659090254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,512,0.014840889308187695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,2048,0.011937778029176923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,7168,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,512,0.007882666256692674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,8192,0.037476443582110934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,3072,0.04277599851290385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,1024,0.33923912048339844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,64,0.02438755498992072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,12288,0.8701066441006131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,3584,0.03437422381507026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,1536,0.08199289109971789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,64,0.003726222034957674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,16384,0.15763289398617214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,5120,0.016521778371598985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,3584,0.9398302502102323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,512,0.009528888596428765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,51200,0.2925288942125109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,8192,0.07374844286176893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,1024,0.005520888914664586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,128,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,2048,0.021927111678653296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,512,0.03700800074471368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,12288,0.012266666524940066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,16384,0.024859555893474158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,128,0.004753777964247597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,16384,0.24009243647257486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,256,0.13670844501919216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,4096,0.01588711142539978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,3584,0.12010667059156631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,1024,0.026211554805437725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,128,0.0035537779331207275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,65536,0.06723822487725152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,3584,0.013605333036846586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,10240,0.024904888537194993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,51200,0.06522755490409003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,10240,0.3301324314541287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,6144,0.08493510882059734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,1536,0.007521777517265744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,128,0.018069333500332303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,512,0.015879111157523263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,1536,0.017649778061442904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,8192,0.05008088880115085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,2048,0.005120000077618493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,2048,0.020986666282018025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,2048,0.07949155569076538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,64,0.006803555621041193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,3584,0.008958222137557136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,51200,0.049667557080586754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,4096,0.015218666858143277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,12288,0.11279200183020698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,2048,0.0068400001360310455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,768,0.01627288924323188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,3072,0.010031110710567897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,4096,0.11010489198896621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,2048,0.007506666911972894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,32,0.002512000087234709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,12288,0.0835919976234436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,1024,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,1024,0.005686222265164058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,12288,5.309622022840712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,3072,0.1165706713994344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,4096,0.008371555142932469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,7168,0.02401688860522376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,512,0.022648000054889258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,8192,0.007259555160999298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,3072,0.19796089331309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,768,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,1536,0.006331555545330048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,2048,0.006534222099516127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,12288,0.03941422369745042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,768,0.16578577624426946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,6144,0.20113600624932182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,12288,0.29021867116292316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,4096,0.00665866666369968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,8192,0.039840890301598444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,16384,9.822840372721354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,2560,0.03669688767857022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,16384,0.05199199914932251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,10240,0.05977333254284329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,2560,0.08158133427302043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,10240,0.018774222996499803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,7168,0.010790222220950656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,2048,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,2560,0.022407111194398668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,3072,0.008286221987671321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,51200,0.8320968945821127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,4096,0.013910222384664746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,3072,0.012347555822796293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,8192,0.021115556359291077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,64,0.016535111599498324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,10240,1.1954613791571724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,51200,0.9069911109076606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,51200,0.8804924223158094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,7168,0.13571377595265707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,7168,0.0170160002178616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,65536,1.5018604066636827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,2560,0.04933155576388041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,2560,0.3016435570187039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,10240,0.0070568886068132185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,3072,0.27962666087680393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,7168,0.01148000028398302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,7168,0.11715822749667698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,6144,0.020922667450375024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,4096,0.7586160235934787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,7168,0.07213244173261854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,16384,1.4053813086615667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,12288,0.08815644184748332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,4096,0.01259022206068039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,4096,0.008368000388145447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,10240,0.08744088808695476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,3584,1.9883467356363933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,32,0.003941333128346337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,64,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,32,0.03630400035116408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,256,0.003768889026509391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,5120,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,8192,0.03288799855444167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,6144,0.006060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,16384,0.05871466795603434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,128,0.00868444475862715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,1024,0.006700444552633498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,5120,0.16839555899302164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,3584,0.011707555916574268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,256,0.005440000030729506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,65536,0.39541512065463597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,2560,0.021827555365032617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,51200,0.1973386605580648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,7168,0.015634665886561077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,6144,0.2202062209447225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,256,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,2048,0.2404551241132948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,10240,0.0144213338692983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,7168,0.027338667048348322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,3584,0.010284444524182213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,1536,0.007547555698288812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,128,0.005025777965784073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,3584,0.018636445204416912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,256,0.006621333460013072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,16384,1.6418000327216253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,3072,0.02641600039270189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,1536,0.04552177919281853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,10240,1.4283342361450195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,12288,0.031982223192850746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,1536,0.018551111221313477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,1536,0.30792710516187877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,256,0.004378666480382283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,7168,0.04291199975543552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,2560,0.057631108495924205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,7168,0.021193777521451313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,12288,0.03232533401913113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,3584,0.07887999878989325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,5120,0.03468799922201369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,51200,0.08464088704850938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,3584,0.027616888284683228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,8192,0.07701510853237577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,6144,3.6811146206325955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,768,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,4096,0.3789369000328912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,7168,0.1295902199215359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,32,0.006017777654859755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,8192,0.011744889120260874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,51200,0.031209776798884075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,1024,0.004120888809363048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,1536,0.056814220216539174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,10240,0.11995022826724583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,10240,0.05147555470466614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,128,0.013059555656380124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,1536,0.12793155511220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,3072,0.25845156775580513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,2560,0.008840888738632202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,1024,0.00796977761718962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,7168,0.10542133119371201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,2048,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,128,0.020850666695170935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,1024,0.005026666654480828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,10240,3.0064656999376087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,8192,0.35832622316148544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,64,0.005839111076460944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,2048,0.010099555883142684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,1024,0.012702222499582501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,5120,0.28880267673068577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,2048,0.03489511211713155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,12288,0.021312889125612047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,65536,0.067193779680464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,512,0.028655999236636694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,16384,0.2832124498155382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,16384,0.48644622166951496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,16384,0.08983733256657918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,8192,0.014347554908858405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,32,0.003583111282851961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,1024,0.010747555229398938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,1536,0.06371644470426771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,7168,0.024200000696712073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,12288,0.037759111987219915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,2560,0.06780978043874104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,5120,0.041013333532545306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,8192,0.07634399996863471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,6144,0.028654221031400893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,10240,1.0071261723836262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,2048,0.010020444790522257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,32,0.02017333275742001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,3072,0.02444800072246128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,256,0.003645333151022593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,2048,0.012015110916561551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,5120,0.22531912061903211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,2560,0.2404515478346083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,768,0.013940445250935025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,2560,0.024917332662476435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,3072,1.0460009045071073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,51200,0.533149348364936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,2048,0.08866222037209405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,2560,0.03052088949415419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,51200,3.234276453653971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,6144,0.01427111029624939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,7168,0.008023111356629265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,2560,0.008044444852405125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,1536,0.011568889021873474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,3584,0.06771555874082777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,256,0.004498666773239772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,256,0.003855111284388436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,128,0.005408888889683618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,4096,0.046851555506388344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,512,0.011676444775528379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,2048,0.2375324567159017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,768,0.00369422220521503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,65536,0.06854755348629422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,3584,0.011360000405046674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,1536,0.005068444543414646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,6144,0.8828924496968588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,8192,0.09102755784988403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,10240,0.02970044480429755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,6144,0.021634666456116572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,65536,0.09042488866382176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,32,0.0033484444850020935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,7168,0.0618471105893453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,65536,0.08298222223917644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,4096,0.02572622232966953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,6144,0.01681688924630483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,512,0.008369777765538957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,10240,0.04874933428234524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,51200,2.477932400173611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,1536,0.04599644409285652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,8192,0.24288090070088705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,768,0.02090311050415039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,256,0.05302133493953281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,6144,0.013148444394270578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,65536,1.4614666832817926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,4096,0.019152000546455383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,64,0.005049777941571342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,768,0.005271111097600725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,2048,0.010623111493057676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,768,0.007717333734035492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,10240,0.675606197781033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,128,0.007198221981525421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,3072,0.0574586656358507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,32,0.01884977850649092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,256,0.005456888841258154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,6144,0.29071111149258083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,7168,0.15751200252109104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,1536,0.05281866590181986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,65536,0.34834665722317165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,768,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,8192,0.011854222251309289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,128,0.005299555758635203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,1536,0.013506666653686099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,2560,0.7307342423333062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,5120,0.19679999351501465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,10240,1.3998355865478516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,1024,0.00629066675901413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,8192,0.02479022244612376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,51200,0.2975200017293294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,128,0.0027173333283927706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,6144,0.27752712037828237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,2048,0.014799111419253878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,3072,0.09031822284062703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,51200,0.8128346867031522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,3584,0.01789155602455139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,16384,0.08051377534866333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,8192,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,5120,0.011949333051840464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,256,0.012430222498046027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,4096,0.1391528844833374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,768,0.04109955496258206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,768,0.05831466780768501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,1024,0.004415110995372136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,65536,1.1498710844251845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,512,0.005614222337802251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,10240,0.013688888814714221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,3584,0.018955555227067735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,3584,0.008882666627566019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,4096,0.0129120002190272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,32,0.00638755535085996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,128,0.015452444553375244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,2560,0.16698755158318412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,768,0.0738453335232205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,32,0.0064586665895250105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,8192,3.7264205084906687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,16384,0.06774844725926717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,6144,0.010251555177900527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,65536,0.10570489035712348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,16384,0.25929511917961967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,4096,0.1759831110636393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,3072,0.01071466671095954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,3584,0.021543999512990315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,1024,0.00499911109606425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,10240,0.9640782674153646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,51200,0.23019822438557944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,51200,0.07169155279795329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,32,0.0034017778105205963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,65536,0.2553679943084717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,1536,0.008646222452322641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,65536,0.17740089363522002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,512,0.003636444608370463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,12288,0.021095999413066443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,512,0.026879999372694228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,1024,0.01034755590889189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,128,0.0053084443012873335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,128,0.017318222257826064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,7168,0.05654578076468574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,1536,0.009654222263230218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,16384,0.10301066769493951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,4096,0.06859288613001506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,7168,0.01463822192615933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,32,0.0052897776994440295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,768,0.005479111025730769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,16384,0.05551911062664456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,16384,0.1597928868399726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,3072,0.00742133292886946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,128,0.004971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,12288,0.07028622097439237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,10240,0.13412533866034612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,65536,0.200455109278361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,3072,0.10971555444929336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,51200,0.07546399699317084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,128,0.0810462236404419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,65536,0.07976355817582872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,1536,0.03278844555219015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,3072,0.010608889162540436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,3584,1.249396430121528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,4096,0.0481697784529792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,4096,0.2081448899375068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,8192,0.019681778219011094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,51200,4.445599026150174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,1024,0.03358666764365302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,10240,0.022261333134439256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,1024,0.009111111362775167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,51200,0.6955377790662977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,32,0.019183110859658983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,5120,2.1196791330973306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,12288,0.9715937508477105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,128,0.008372444245550368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,512,0.004555555681387584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,768,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,512,0.06951022148132324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,3584,0.016061334146393668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,51200,0.06188266807132297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,6144,0.06008266740375095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,1024,0.00629066675901413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,16384,0.026484444737434387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,51200,0.6415520244174534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,12288,0.05537422166930305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,65536,0.46535645590888125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,512,0.09608888626098633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,4096,0.021535111798180476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,512,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,768,0.00872444444232517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,32,0.003071111109521654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,7168,0.23523465792338052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,6144,0.11703999837239583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,7168,0.010553778045707278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,2048,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,4096,0.011879111329714457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,65536,1.279831992255317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,12288,1.048653284708659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,5120,0.07641955878999498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,1536,0.016680000556839835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,7168,0.06385688649283515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,768,0.2806737687852648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,51200,0.06306933032141791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,64,0.00498844434817632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,7168,0.04855555627081129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,768,0.005756444401211209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,10240,0.018408889571825664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,2560,0.028853333658642236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,256,0.09946133030785455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,128,0.00692622239391009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,10240,0.014527110589875115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,2048,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,768,0.01902311046918233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,6144,0.01815466748343574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,65536,1.424658669365777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,3072,0.05944533480538262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,768,0.18883555465274385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,51200,0.177818669213189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,10240,0.014715555641386243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,51200,0.047445333666271634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,16384,0.030151999659008447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,10240,0.013635555903116861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,8192,0.736128012339274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,128,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,4096,0.03251733382542928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,3072,0.021830222672886316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,16384,0.28519021140204537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,10240,0.05405155486530728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,5120,0.032680888970692955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,5120,0.25242755148145885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,8192,0.027375110321574744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,6144,0.009902222288979424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,16384,0.038682666089799665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,2048,0.01941155559486813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,3584,0.012372444073359171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,2048,0.00756444450881746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,65536,0.4600426885816786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,10240,0.48559289508395725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,10240,0.01739288866519928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,3584,0.013246222502655454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,8192,0.022250667214393616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,256,0.0032426667296224167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,65536,0.1945706605911255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,3072,0.15904888841840956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,16384,0.020620443754725985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,1024,0.0041422223051389056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,10240,0.010738666686746808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,32,0.008750222623348236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,1536,0.054048889213138156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,7168,0.15190311272939047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,10240,0.16205778386857775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,16384,0.06472533278995089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,128,0.002983999955985281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,7168,1.0356000264485676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,2560,0.1253839996125963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,128,0.008117333054542542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,4096,0.023575110567940608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,128,0.009684444301658207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,2560,0.009540444446934594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,51200,0.27231287956237793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,6144,0.01143200033240848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,65536,1.0924835205078125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,3584,0.08909244669808282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,1536,0.0058444444504049085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,128,0.007618666523032718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,256,0.004913777940803104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,128,0.0107360002067354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,3072,0.5454542371961806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,12288,0.08957511186599731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,51200,0.05596444341871473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,256,0.0068400001360310455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,5120,0.024466666910383437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,7168,0.04359644320276049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,768,0.004825777891609403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,51200,6.82586669921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,6144,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,7168,0.3274986743927002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,4096,0.16997778415679932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,10240,0.5264977878994411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,256,0.007792889244026608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,1024,0.003983111017280155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,2560,0.25892355706956655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,16384,0.3834373421139187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,512,0.013739555246300168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,256,0.00889511075284746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,2048,0.010874666273593903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,512,0.02179466684659322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,12288,1.0026782353719075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,12288,0.03388088941574097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,1024,0.00509333320789867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,2048,0.0052897776994440295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,5120,0.029601779248979356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,1536,0.7077191140916613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,4096,0.2108231120639377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,8192,0.04032355546951294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,64,0.007300444775157505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,64,0.006815111057625876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,2560,0.04419733418358696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,1024,0.009492444495360056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,4096,0.008454221818182204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,768,0.007473777565691207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,3584,0.03417244553565979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,3072,0.026688888669013977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,1024,0.017097777790493436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,4096,0.07447466585371229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,256,0.0032355555643637977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,4096,0.01036977767944336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,7168,0.2081306642956204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,256,0.003173333282272021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,10240,3.882782406277127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,5120,0.04892533355289035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,8192,0.08774133523305257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,64,0.014852444330851236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,128,0.0058648888435628675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,128,0.008022222254011366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,5120,1.5524347093370225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,16384,0.19938488801320395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,2560,0.07394933038287692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,12288,0.07525066534678142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,2560,0.028625776370366413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,512,0.005579555614127054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,65536,2.02767456902398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,2560,0.011789333489206104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,7168,0.02979733215437995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,16384,0.7859288851420084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,8192,0.03384888834423489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,6144,0.022978666755888198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,6144,0.017607111069891188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,768,0.030641777647866145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,256,0.004536889079544279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,1024,0.008766222331258986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,7168,0.006621333460013072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,1024,0.04914933443069458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,3584,0.07901422182718913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,8192,0.37340709898206925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,7168,0.10629422134823269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,512,0.05102666550212436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,64,0.03903466794225905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,12288,0.26133778360154897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,16384,0.02720444401105245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,51200,1.0260559717814128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,128,0.002959111084540685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,4096,0.01959377858373854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,3072,0.04345688886112637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,2560,0.014949333336618213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,16384,0.09395466910468207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,1024,0.03295022249221802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,512,0.003584889074166616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,10240,0.014056889547242058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,3072,0.04177155428462558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,512,0.00554577757914861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,3584,0.20059822665320504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,10240,0.19538044929504395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,7168,0.010680889089902243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,3584,0.022663111488024395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,2560,0.2984444565243191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,16384,0.13290044996473524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,768,0.00647288892004225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,65536,0.25263910823398167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,3072,0.009532444179058075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,65536,0.014460444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,4096,0.061439110173119434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,2048,0.00926577796538671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,16384,0.02404888967672984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,16384,0.31380976570977104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,8192,0.020265777905782063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,128,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,1536,0.02998222245110406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,128,0.003069333318206999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,1536,0.006331555545330048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,6144,0.037689778539869524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,4096,0.010298666854699453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,7168,0.05603288941913181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,3584,0.1191564400990804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,64,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,1024,0.0118231111102634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,10240,26.19088406032986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,10240,0.4967350959777832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,64,0.016594666573736403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,3072,0.10485600100623237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,51200,0.022840888963805303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,32,0.016979555288950603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,12288,0.07643022139867146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,3584,0.25910311275058323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,2048,0.011240888800885944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,512,0.003273777870668305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,768,0.022623111804326374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,64,0.19479910532633463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,51200,0.053244445059034556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,12288,0.01625866691271464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,256,0.005429333282841577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,16384,0.05021333363321093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,256,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,10240,0.9512515597873263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,6144,0.4224835501776801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,768,0.010738666686746808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,3584,0.015171556009186639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,1024,0.005716444303592046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,51200,0.08633955319722493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,512,0.003995555556482739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,51200,0.05297155512703789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,512,0.010256888965765635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,512,0.019683554768562317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,2048,0.028241776757770117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,4096,0.8467395570543078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,4096,0.010393777655230628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,1536,0.010692444112565784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,2048,0.01202666676706738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,65536,0.01699199941423204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,3584,0.07120355632570055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,6144,0.20909867021772596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,4096,0.0107360002067354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,3072,0.007768889268239339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,8192,0.35522132449679905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,768,0.005740444279379315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,7168,0.13418489032321507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,12288,0.02036177780893114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,32,0.10824888944625854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,2048,0.010875555376211802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,5120,0.16090667247772217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,512,0.0053377776510185665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,8192,1.2669750849405925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,4096,0.07741244633992513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,2048,0.03391644358634949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,6144,0.34644799762301975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,65536,0.3574097686343723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,51200,0.08816088570488824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,51200,0.09235999981562297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,128,0.0027955555253558685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,2560,0.027829332484139338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,8192,1.5526826646592882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,128,0.0038319999972979226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,2560,0.011253332926167382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,2048,0.0695671108033922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,64,0.003084444544381566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,16384,0.047246221039030284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,2560,0.0602417786916097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,65536,0.14634577433268228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,8192,0.025071110990312364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,51200,1.6415600246853297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,3072,0.04823288983768887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,7168,0.4075831042395697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,128,0.003257777748836411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,65536,0.5835679901970757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,256,0.003212444484233856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,1024,0.00630044440428416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,12288,0.3402542273203532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,5120,0.008982222113344405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,3072,0.010671111444632212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,51200,0.6325048870510525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,3072,0.044643554422590465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,7168,0.016226665841208566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,256,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,3584,0.049477332168155246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,6144,0.05075110991795858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,5120,0.09172533618079291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,4096,0.05301510956552294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,3072,0.007975111405054728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,12288,0.1465653313530816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,10240,0.015929778416951496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,2560,0.01791822248035007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,7168,0.012317332956526013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,2560,0.005656888915432825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,8192,0.0358240008354187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,128,0.03250133328967624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,6144,0.02490666674243079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,512,0.006513777706358168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,256,0.007034666836261749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,8192,0.053921778996785484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,10240,2.829716364542643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,10240,0.03421511252721151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,64,0.005046222358942032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,128,0.01681866745154063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,51200,0.07644533448749118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,16384,0.2479591104719374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,6144,0.24342666731940377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,51200,3.2725022633870444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,2560,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,8192,0.05552000138494703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,32,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,128,0.014703111516104804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,3072,0.09101066986719768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,65536,0.24926222695244682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,16384,0.2026275528801812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,8192,0.8546551068623861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,512,0.004607111215591431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,512,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,16384,0.031239999665154353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,128,0.017039111918873258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,65536,4.458988613552518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,51200,7.157042609320746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,12288,0.3346302244398329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,6144,0.0181031111213896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,2560,0.008791999684439765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,2560,0.013206221991115145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,6144,0.035990221632851496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,768,0.01788444485929277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,5120,0.016712889075279236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,8192,0.045724444919162326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,3584,0.010584000084135266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,64,0.0052639999323421055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,128,0.0244186669588089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,256,0.025619554850790236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,6144,0.030598223209381104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,10240,0.055656890074412026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,16384,2.4783787197536893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,3584,0.01696888936890496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,128,0.002883555574549569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,64,0.02732444471783108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,5120,0.01204533295498954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,12288,0.03616088959905837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,6144,0.9556542502509223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,7168,0.030051555898454454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,5120,0.07324444585376315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,128,0.003032888803217146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,4096,0.03682133224275377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,4096,0.0402471125125885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,64,0.005172444300519096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,65536,0.061662223603990346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,6144,0.011815110842386881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,1024,0.21952533721923828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,128,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,12288,0.012857777376969656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,10240,0.18702844778696695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,7168,0.034938666555616595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,7168,0.08685866991678874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,10240,0.039847112364239164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,7168,0.09666133589214748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,10240,0.38184089130825466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,32,0.0030177777840031516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,768,0.006696888970004187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,3072,0.08412266439861721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,8192,0.022742221752802532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,16384,0.32653244336446124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,51200,4.179790920681424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,4096,0.08813244766659206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,4096,0.26835110452440053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,51200,0.5543431176079644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,4096,0.03432000014517043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,4096,0.006637333167923822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,768,0.08738133642408584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,10240,0.09165244632297093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,3584,0.012410666379663678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,128,0.09371466769112481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,8192,0.03191733360290527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,51200,0.9527191585964627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,4096,1.0978915956285265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,16384,0.046330667204327054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,4096,0.6961386468675402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,768,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,5120,0.2735297679901123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,5120,0.030398223135206435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,64,0.04863288998603821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,256,0.004269333349333869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,256,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,7168,0.5915093421936035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,512,0.007768889268239339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,1024,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,256,0.013095999757448832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,65536,2.789804458618164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,16384,0.5595128801133897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,64,0.003181333343187968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,4096,0.008599999878141616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,6144,0.226149320602417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,128,0.004466666529575984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,2048,0.01664622293578254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,8192,0.06595377789603339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,768,0.012111110819710625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,6144,0.06517955329683092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,4096,0.18725421693589953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,2560,0.15474843978881836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,51200,0.15849155849880642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,768,0.26469243897332084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,6144,0.014473777678277759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,768,0.04064355625046624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,65536,0.9127226935492622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,1536,0.006568000134494569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,1024,0.0070293330483966405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,1536,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,1536,0.010480889015727572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,512,0.0032986667421129015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,65536,0.17528533935546875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,4096,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,65536,0.9159688949584961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,1536,0.12306933932834202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,32,0.012335110869672565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,8192,0.031238221459918555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,12288,0.026100445124838088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,1536,0.008993777963850234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,10240,0.40306666162278915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,65536,0.17596977286868623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,5120,0.15863822566138372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,1024,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,7168,0.18764977984958223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,1536,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,128,0.21092622809939912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,7168,0.008449777960777283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,1024,0.01180622229973475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,6144,0.013628444737858243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,256,0.012677333421177335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,7168,0.11345511012607151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,512,0.008423999779754216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,6144,0.1338311168882582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,768,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,4096,0.013978666729397245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,256,0.004452444612979889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,512,0.005608888963858287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,2560,0.07230577866236369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,3072,0.017071111334694754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,4096,0.061666667461395264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,12288,0.09628355503082275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,64,0.007431999676757389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,128,0.01682844426896837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,10240,0.15761866834428576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,6144,0.013551111022631327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,16384,0.007823111282454597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,512,0.013438222308953604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,2560,0.009519110951158736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,32,0.024127110838890076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,64,0.03380799955791897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,3584,0.031002667215135362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,512,0.004856889032655292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,128,0.18554488817850748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,32,0.026692445079485577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,128,0.007755555212497711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,16384,0.4408453305562337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,1024,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,1024,0.025349333882331848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,10240,0.03950044512748718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,2560,0.37016622225443524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,768,0.007727111379305522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,2560,0.0070186663005087115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,1536,0.05252444412973192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,12288,0.02992977698644002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,51200,0.5478462113274468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,10240,0.056792888376447886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,5120,0.0348693331082662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,5120,0.007344000041484833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,10240,0.030802667140960693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,3584,0.01629244453377194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,3072,0.009150221943855286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,6144,0.010619555910428366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,12288,0.056167999903361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,65536,0.091666665342119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,4096,0.09874577654732598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,4096,2.4430294036865234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,51200,3.2442550659179688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,8192,0.6185359954833984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,8192,0.03029688861634996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,5120,0.02072533302836948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,32,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,2048,0.13184977902306452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,1536,0.11042488945855035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,6144,0.014310222533014087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,128,0.007206222249401941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,3584,0.04770133230421278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,3584,0.1953831116358439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,3072,0.054485334290398486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,2560,0.008761777646011775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,4096,0.05773599942525228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,16384,0.0911182231373257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,2560,0.04724444283379448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,3072,0.06545244322882758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,1536,0.01975733372900221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,1024,0.01075644459989336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,512,0.010846222440401712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,4096,0.022429333792792425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,10240,0.01444533301724328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,5120,0.49726221296522355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,6144,0.15012444390190974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,10240,0.35779378149244523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,16384,0.1289786630206638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,6144,0.02629511058330536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,2560,0.007747555772463481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,32,0.011555555793974133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,16384,0.2418959935506185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,1536,0.008430221842394935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,16384,0.10399111111958821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,3072,0.010797333386209277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,3584,1.2913102043999565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,768,0.00963644435008367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,128,0.003371555565132035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,512,0.018221333622932434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,3072,1.0309840308295357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,7168,0.01867733399073283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,6144,0.11481332778930664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,7168,0.019174221489164565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,4096,0.012310222619109683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,2048,0.02084355552991231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,6144,0.02306844459639655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,3072,0.010216889282067617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,1024,0.06560355424880981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,512,3.9905849032931857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,1024,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,512,0.05138488941722446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,128,0.055368890364964805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,51200,0.16861511601342094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,768,0.015203555425008139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,1024,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,16384,0.02641600039270189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,8192,0.029232889413833618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,2048,0.05902666515774197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,4096,0.01038577738735411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,512,0.0042204442951414315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,2048,0.01663644446267022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,3072,0.011076444553004371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,256,0.01610488858487871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,16384,0.10777066813574897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,256,0.20077599419487846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,1536,0.01997155613369412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,128,0.0037902221083641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,2560,0.006142222219043308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,8192,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,1536,0.014111111561457316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,2560,0.015160888433456421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,51200,0.9015688366360135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,16384,1.3273341920640733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,64,0.008630221916569604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,10240,0.052941332260767617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,1536,0.006773333168692059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,1024,0.01259200026591619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,256,0.0988337794939677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,1536,0.1427315606011285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,2560,0.014748444159825643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,3072,0.011355555719799466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,51200,1.2133777406480577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,1024,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,1536,0.038485334979163274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,16384,0.02115733259254032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,64,0.013339555925793119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,16384,0.11655733320448135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,2560,0.11348977353837754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,256,0.006648889018429651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,1536,0.018925334016482036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,2560,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,2048,0.008727110922336578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,4096,0.047896001074049205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,2560,0.022304000125990972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,65536,3.1708666483561196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,64,0.003370666669474708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,1024,0.007479111353556315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,512,0.013016888664828407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,3072,0.016720000240537856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,3584,0.014529777897728814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,1024,0.0225262227985594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,65536,0.22747556368509927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,256,0.017664889494578045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,3584,0.04350577791531881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,3072,0.2404773235321045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,768,0.00498933345079422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,128,0.004419555680619346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,1536,0.07697421974605985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,3072,0.01863911085658603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,8192,0.06291733185450236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,16384,1.9688932630750868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,16384,0.02029333346419864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,3072,0.009944889280531142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,8192,0.5403235753377279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,6144,0.1258880032433404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,12288,0.7939119868808322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,10240,0.026455110973782007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,3072,0.052503999736573964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,6144,0.12600444422827825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,7168,0.008757332960764566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,51200,0.024754666619830664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,256,0.012225777738624148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,1024,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,32,0.6805120044284396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,32,0.005148444324731827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,16384,0.00980000032318963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,1024,0.006181333214044571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,128,0.009135111338562435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,768,0.008162666526105668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,4096,0.01439377831088172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,8192,0.2759573459625244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,5120,0.09793688853581746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,10240,0.767815113067627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,768,0.006953777538405524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,32,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,256,0.005099555684460534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,12288,0.14555733733707005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,1024,0.009476444787449306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,5120,0.034380445877710976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,16384,0.5371848742167155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,3584,0.016920000314712524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,2048,0.007714666426181793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,3584,0.020039111375808716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,10240,0.02863466739654541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,65536,0.41132177246941465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,6144,0.3660399913787842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,32,0.007000889215204451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,64,0.0064604443808396654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,65536,0.12205778227912055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,8192,0.04009688893953959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,12288,0.05793244308895535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,8192,0.12220177385542129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,5120,0.010444444086816577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,51200,0.2610417736901177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,768,0.026383999321195815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,10240,0.2753946781158447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,512,0.032792889409595065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,1024,0.013576000101036496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,768,0.03100444542037116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,1024,0.010017777482668558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,12288,0.7855324215359158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,128,0.00480000012450748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,1536,0.04251911242802938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,7168,0.014532445205582513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,6144,0.01035733355416192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,7168,0.030101332399580214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,2048,0.025840888420740765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,128,0.009134222235944536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,512,0.012650666965378655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,64,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,512,0.025527111358112756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,16384,0.08324978086683485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,5120,0.010354667074150508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,7168,0.008585777547624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,32,0.008960000342792934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,6144,0.47920624415079754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,12288,0.04694044589996338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,7168,0.24493334028455946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,256,0.0058417779703934984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,12288,0.24674312273661295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,3072,0.007182222273614671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,10240,0.018557333283954196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,128,0.0064186664919058485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,10240,0.03501510951254103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,51200,0.07086400190989177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,5120,0.15543556213378906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,3584,0.028254220883051556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,2560,0.43386133511861164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,65536,0.17928622828589547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,8192,0.4517582257588704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,1024,0.025810667210155066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,5120,0.012168000141779581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,7168,0.022673777408070032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,512,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,2048,0.06821866830190022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,512,0.010827555424637265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,1536,0.020136000381575692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,8192,0.018617777360810172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,2560,1.3416115442911785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,6144,0.015588444140222339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,768,0.010510221951537661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,64,0.02309422194957733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,65536,0.47687647077772355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,65536,0.06915466652976142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,3072,0.008174222376611497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,3584,0.008489777644475302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,12288,0.13126488526662192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,7168,0.07856533262464735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,1536,0.010631110933091907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,51200,0.21980533334943983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,3584,0.061518225404951304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,10240,0.019488000207477145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,3584,0.03788177834616767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,12288,0.04066933194796244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,3584,0.026159110996458266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,65536,0.26965777079264325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,512,0.003539555602603488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,1536,0.14459555678897434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,2048,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,6144,0.33632977803548175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,16384,0.05651999844445122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,7168,0.023007111416922674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,32,0.005453333258628845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,64,0.015622221761279635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,128,0.004412444515360726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,128,0.005919999960396025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,10240,0.08102399773067899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,1024,0.008808000220192803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,16384,0.05006577902370029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,2560,0.0074933336840735535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,768,0.04860888918240865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,64,0.004607999904288186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,768,0.04020444552103678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,5120,0.1556817822986179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,7168,0.012249777714411417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,768,0.1560693316989475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,5120,0.2146159940295749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,6144,0.021822222405009802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,3584,0.04563022322124905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,256,0.0114720000161065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,1024,0.004785777793990241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,32,0.04537422127193875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,768,0.01201777739657296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,10240,0.00871022211180793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,64,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,2560,0.04141155547565884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,512,0.01402933398882548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,65536,1.6274799770779078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,65536,1.7274328867594402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,2048,0.041639112763934664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,6144,0.10264711247550116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,10240,0.10478578011194865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,16384,0.02462399999300639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,5120,0.011867555479208628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,768,0.053616000546349414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,256,0.01629155543115404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,10240,5.095792134602864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,64,0.11259022023942734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,2048,0.008921778036488427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,6144,0.04399999976158142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,768,0.006090666684839461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,2048,0.13117599487304688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,51200,0.2733173370361328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,65536,3.205848905775282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,768,0.0343502230114407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,51200,0.9496986601087781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,10240,0.07436355617311266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,65536,7.2820078531901045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,128,0.011864888999197217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,10240,0.030344890223609075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,1024,0.009742221898502773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,1536,0.014933332800865173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,4096,0.008803555534945594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,3072,0.012959111067983838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,4096,0.016334222422705758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,3584,0.0169813334941864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,1024,0.005968889014588461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,16384,0.0861004458533393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,5120,0.018765333626005385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,256,0.0033440000067154565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,16384,0.2376640107896593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,32,0.012048888537618848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,4096,0.014242667290899487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,768,0.005576000031497743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,51200,0.6443262100219727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,4096,0.1375653346379598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,4096,0.01734133395883772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,16384,0.026520000563727483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,7168,0.006586666736337874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,4096,0.010701333483060202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,768,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,64,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,6144,0.34459378984239364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,256,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,2048,0.005388444496525659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,2560,0.011176889141400656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,12288,0.030061334371566772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,2560,0.014661333627170987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,2560,0.09369333585103352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,51200,2.4070311652289496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,4096,0.0679253339767456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,32,0.004085333396991094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,65536,4.845176696777344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,7168,0.03145688772201538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,7168,0.05308177736070421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,128,0.008627555436558193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,3584,0.015487111277050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,512,0.007055111229419708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,5120,0.029159112109078303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,65536,0.25169067912631565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,128,0.004134222037262387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,65536,0.06621155473921034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,16384,0.33417333496941465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,3072,0.011977777712874942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,7168,0.020557334025700886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,10240,0.026717333330048457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,1024,0.05135111014048258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,3584,0.052319112751219005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,65536,0.539522647857666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,4096,0.009443555441167619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,2560,0.03254222207599216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,128,0.0031439999325407874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,3072,0.16066222720676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,1536,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,3584,0.06405599912007649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,128,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,32,0.016892444756295946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,768,0.006089777582221561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,768,0.01147911118136512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,65536,3.2119839986165366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,32,0.046294222275416054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,512,0.008850666383902231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,1536,0.026909333136346605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,3072,0.14083377520243326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,1536,0.02328799996111128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,6144,0.13062666522132024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,2048,0.012454222473833295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,3584,0.013588444226317935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,16384,0.14461689525180393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,3072,0.01607466737429301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,3072,0.06483555502361722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,128,0.00498933345079422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,3072,0.008587555752860175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,51200,0.053704000181621976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,512,0.22317067782084146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,256,0.003721777763631609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,5120,0.011456888582971362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,5120,0.2122337818145752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,12288,4.296918233235677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,2560,0.1985173357857598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,768,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,51200,2.163727018568251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,512,0.8704462051391602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,128,0.0034426665968365143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,3584,0.01828888886504703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,10240,0.026299556096394856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,512,0.046703110138575234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,128,0.005440000030729506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,256,0.008466666771305932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,128,0.00517155561182234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,64,0.0033297776761982176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,2560,0.007344000041484833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,10240,0.26688265800476074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,8192,0.05388444330957201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,16384,4.413135104709201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,10240,2.8378630744086375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,768,0.015957333975368075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,128,0.0031804444475306403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,128,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,2560,0.036502222220102944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,51200,2.813067542182075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,1536,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,64,0.0028497777465316984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,6144,0.611629327138265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,3072,0.01347644461525811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,8192,0.5018026563856337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,512,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,64,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,768,0.006351111249791251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,6144,0.01350311107105679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,1024,0.007246221933099959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,51200,0.2739200062221951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,51200,0.053489777776930064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,12288,0.050195554892222084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,7168,0.2588204542795817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,512,0.009505777723259395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,7168,0.0365937782658471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,2560,0.028556442923016016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,4096,0.05236444539493985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,1536,0.008247111406591203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,4096,0.1321235497792562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,5120,0.010684444672531553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,12288,0.047468443711598717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,5120,0.13369866212209067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,1536,0.004304000073009067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,128,0.037793778710895114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,65536,0.06217777729034424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,32,0.004295111116435793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,256,0.0030746666921509635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,256,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,10240,0.11949066321055095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,32,0.0032319999817344877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,256,0.0069857777820693115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,512,0.04650133185916477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,5120,0.02756888833310869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,768,0.007330666813585494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,256,0.004558222161398994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,2048,0.006650666809744305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,3072,0.01700444519519806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,16384,0.16646756066216364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,256,0.0341306659910414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,3072,0.023784889115227595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,2048,0.00489155575633049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,7168,0.8957608540852865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,3072,0.010148444109492833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,2560,0.034281777011023626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,1024,0.005471111171775394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,5120,0.020985777179400127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,32,0.002990222225586573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,512,0.008385777473449707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,51200,0.6414888699849447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,4096,0.06610221995247735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,1536,0.07422133286794026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,8192,0.1580915583504571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,8192,0.22935822274949816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,51200,0.41628087891472715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,1536,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,51200,0.3200959894392225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,7168,0.2187324497434828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,512,0.020917332834667627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,32,0.00452977791428566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,51200,0.06907200151019625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,7168,0.06562666760550605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,256,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,4096,0.488700442843967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,3584,0.043747554222742714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,1024,0.009306666751702627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,768,0.0058542220956749385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,3584,0.21781955824957955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,3072,0.01921688848071628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,2560,0.0214711113108529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,3584,0.01088711122671763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,32,0.007173333730962541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,16384,0.024381332927280005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,12288,0.2530826727549235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,4096,0.010141332944234213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,256,0.00443111111720403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,2560,0.06142489115397135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,512,0.012039999994966718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,5120,0.02427733275625441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,3584,0.038570665650897555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,128,0.01148000028398302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,65536,0.2681173218621148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,2048,0.007063999772071838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,5120,0.05417155557208591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,256,0.0070568886068132185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,64,0.0052382221652401825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,4096,0.01053244454993142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,3584,0.010323555933104621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,12288,0.2621440092722575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,2048,0.12237955464257134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,65536,2.0214879777696395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,16384,0.08579288588629828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,10240,0.037913779417673744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,10240,0.1336302227444119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,768,0.026711111267407734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,7168,0.0743395553694831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,51200,19.988008287217884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,51200,0.20332088735368517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,65536,0.3554826577504476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,128,0.004276444514592488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,2048,0.007717333734035492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,768,0.027479110492600337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,1024,0.006502222269773483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,128,0.00396799999806616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,64,0.003228444399105178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,1536,0.5784248775906032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,2560,0.007116444408893585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,512,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,2048,0.07993511358896892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,3584,0.010474666953086853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,1536,0.09841955370373195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,4096,0.01420088940196567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,65536,0.5880142317877876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,1536,0.024311999479929607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,32,0.00572622236278322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,2560,0.22280889087253145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,3072,0.045447998576694064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,1024,0.016826666063732572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,3584,0.025368889172871906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,3072,0.2398311032189263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,7168,0.17244444953070748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,512,0.00811822215716044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,12288,0.029986666308508977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,1024,0.012131555212868584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,128,0.022069333328141108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,51200,0.23071021503872344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,1536,0.004816888935036129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,7168,0.21041599909464517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,16384,0.05561511384116279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,128,0.004419555680619346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,1024,0.0052355556852287715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,65536,0.016657777958446078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,3072,0.016737777325842116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,3072,0.06999999947018094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,768,0.05416799916161431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,32,0.007849777738253275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,32,0.010011555420027839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,2560,0.007262222468852997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,2048,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,1536,0.008090666598743862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,5120,0.009393778112199571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,768,0.015504888362354703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,10240,3.628385755750868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,1536,0.03168177935812209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,3072,0.03414400087462531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,128,0.005148444324731827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,8192,1.9699022505018446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,2048,0.008190222084522247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,4096,0.008814222282833522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,7168,0.30231110254923504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,8192,0.29556534025404185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,2560,0.012395555774370829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,768,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,2048,0.009311111436949836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,64,0.027073777384228174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,8192,0.02639466689692603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,32,0.006545777950021956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,6144,0.017652443713612027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,128,0.005047111047638787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,32,0.02484888831774394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,4096,0.014382221632533602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,1536,0.017142222987280954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,5120,0.016202666693263583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,1536,0.006686222222116258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,512,0.007690666450394525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,3584,0.07186044587029351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,5120,1.3895875083075628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,3584,0.011160000330872007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,128,0.004970666848950916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,768,0.06078222062852648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,12288,0.1142115592956543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,1536,0.014732443624072604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,2048,0.06207199891408285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,2048,0.005606222069925732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,51200,0.36190043555365664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,32,0.0035457776652442086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,5120,0.042429334587521024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,4096,0.031311998764673867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,128,0.0028293333533737394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,128,0.003051555612021022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,4096,0.8861484527587891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,6144,0.033667554457982383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,1024,0.3388044569227431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,768,0.011745777395036487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,2560,0.007877333296669854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,3072,0.018719999326599967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,10240,0.015549333559142219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,1536,0.006981333096822103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,5120,0.0728071133295695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,7168,0.02595111065440708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,64,0.004587555511130227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,256,0.00870311094654931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,512,0.029841777351167467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,768,0.004600000050332812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,6144,0.0151591118839052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,65536,0.027890665663613215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,2560,0.018224888377719455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,10240,0.024495111571417913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,6144,0.11458755864037408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,2048,0.11699733469221328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,7168,0.02898755669593811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,4096,0.014937778313954672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,512,0.030339555607901678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,8192,0.39846311675177676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,64,0.007755555212497711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,51200,0.05858400132921007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,12288,0.030822220775816176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,256,0.004851555658711327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,10240,0.044567111465666026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,512,0.030941334035661485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,12288,0.006141333530346553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,768,0.0046844445168972015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,6144,0.010853333605660332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,7168,0.030741333961486816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,3072,0.007394666473070781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,65536,0.06332800123426649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,256,0.008752889103359645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,256,0.007253333098358578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,128,0.004100444416205089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,256,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,3584,0.007763555480374231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,1024,0.008289777570300633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,7168,0.1345235506693522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,128,0.0027351110345787476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,1536,0.00961599995692571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,3072,0.06901244322458903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,2048,0.016082665986484952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,8192,0.46770487891303164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,2560,0.008254222571849823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,10240,1.6246808369954426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,1024,0.011578666667143503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,1024,0.004906666775544484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,128,0.00813422186507119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,16384,0.06462399827109443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,8192,0.08269866969850329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,16384,0.13550933202107748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,256,0.022837332553333704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,10240,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,65536,0.06746221913231744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,12288,0.21574221716986763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,8192,0.0338986681567298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,6144,0.01038044442733129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,768,0.0047395556337303585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,5120,0.010671111444632212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,12288,0.08444178104400635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,3584,0.015772443678643968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,512,0.015029332704014249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,7168,0.23904000388251412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,2048,0.1541937722100152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,5120,0.0702435572942098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,51200,0.2351315551333957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,12288,0.24147023095024955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,5120,0.1375866731007894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,128,0.02128977742460039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,768,0.005619555711746216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,768,0.013388444152143268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,768,0.005180444568395615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,6144,0.024554666545655992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,6144,0.09985688659879897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,2048,0.007896000312434303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,8192,0.015976000163290236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,2048,0.017255110873116385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,1536,0.0070159998204973005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,1536,0.007169778148333232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,1024,3.4656630622016062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,5120,0.012878222597969903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,64,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,5120,0.019388443893856473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,64,0.002520888836847411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,2560,0.048840890328089394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,32,0.003648888733651903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,3072,0.00853066643079122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,65536,0.2637564341227214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,128,0.0032248888164758682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,64,0.02354755500952403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,3072,0.01760088900725047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,256,0.004675555560323927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,65536,0.024957334001859028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,128,0.019849777221679688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,16384,0.12559466891818577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,2560,0.007364444434642792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,3584,0.024482667446136475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,51200,10.401468912760416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,768,0.01737600068251292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,8192,1.5171226925320094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,16384,0.030132446024152968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,256,0.029496000872717962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,12288,0.00701333334048589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,6144,0.073889778720008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,12288,0.021189333664046392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,1536,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,32,0.01806400054030948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,6144,0.024860444996092055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,64,0.005012444323963589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,3072,0.007140444384680853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,51200,0.09133688608805339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,1024,0.006807111203670502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,5120,0.0083155557513237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,512,0.010138666464222802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,32,0.0028328889360030494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,65536,0.5268453492058648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,128,0.02681333323319753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,12288,0.3765093220604791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,64,0.08946044577492608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,2560,0.007647111184067196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,1536,0.009170666337013245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,3584,0.033554666572146945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,8192,0.351748440000746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,768,0.005566222386227713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,51200,0.9520880381266276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,3072,0.15869243939717612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,7168,0.13962933752271864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,3072,0.010099555883142684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,64,0.00583022211988767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,2048,0.008251555263996124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,32,0.006897777732875612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,8192,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,1024,0.02565066681967841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,512,0.044900443818834096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,4096,0.00851466672288047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,6144,0.18716800212860107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,1536,0.010018666585286459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,768,0.017305778132544625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,5120,0.009127111070685917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,65536,6.494166904025608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,5120,0.01904355486234029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,5120,0.06275555822584364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,256,0.0068897778789202375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,2048,1.6487066480848525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,7168,0.020655999581019085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,10240,0.041794667641321816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,8192,0.014485332700941296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,2560,0.03125422199567159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,128,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,1024,0.00685511115524504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,10240,0.022907555103302002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,6144,0.10788355271021526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,12288,0.10247821940316094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,768,0.34076976776123047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,1536,0.0058453331391016645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,512,0.029747555653254192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,10240,0.07399644454320271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,4096,0.16793155670166016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,7168,2.6059822506374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,2048,0.25948090023464626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,256,0.03184977836079068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,2560,0.0169804443915685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,128,0.021681777305073206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,8192,0.02849688794877794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,128,0.01037599974208408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,51200,0.021664889322386846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,3584,0.06589688857396443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,2560,0.016297777493794758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,6144,0.012977778083748288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,768,0.023500444160567388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,1536,0.03791911072201199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,128,0.011683555940786997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,512,0.006124444305896759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,8192,1.228702227274577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,16384,0.07968355549706353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,16384,0.47068267398410374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,12288,0.016919111212094624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,64,0.0033653332955307434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,768,0.004247111164861255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,64,0.0163093325164583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,32,0.007253333098358578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,4096,0.054551111327277295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,3584,0.12146045102013482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,7168,7.681922064887153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,1024,0.01793155570824941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,8192,0.6802426444159614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,10240,0.07928622431225248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,4096,0.021718222233984206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,6144,0.1336559984419081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,8192,0.013443555268976422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,768,0.007899555895063613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,64,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,128,0.04990577697753906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,6144,0.0255057778623369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,4096,0.021720000439220007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,2048,0.015964445140626695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,128,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,16384,0.055620445145501025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,3584,0.009829333258999718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,1536,0.004652444273233414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,4096,0.20442844761742485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,16384,0.6525715721978081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,512,0.005704888867007361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,5120,2.593682607014974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,2560,0.46732621722751194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,65536,7.838999430338542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,1536,0.0258879992696974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,12288,0.15884799427456325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,1536,0.013597333596812354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,256,0.007084444165229797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,1024,0.025757332642873127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,5120,1.412976900736491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,7168,0.02141688929663764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,2048,0.009640889035330879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,256,0.011122666299343109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,2048,0.377987543741862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,64,0.002817777709828483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,65536,0.11592266294691299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,10240,0.00628977765639623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,51200,0.021402666966120403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,512,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,2560,0.06115555763244629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,12288,0.11523377895355225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,1024,0.1170426607131958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,4096,0.0058471109304163195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,32,0.16345422797732884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,5120,0.013743110828929476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,128,0.005323555734422471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,3584,0.011682666838169098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,512,0.03759733173582289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,10240,0.07140977515114678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,128,0.010272000398900773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,51200,0.053936001327302724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,51200,0.18041511376698813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,10240,0.009321778184837764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,5120,0.1664702230029636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,512,0.06570577621459961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,1024,0.019063111808564927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,1024,0.005892444401979446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,2048,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,8192,0.0103511114915212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,16384,0.05001688996950785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,2560,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,8192,0.0413991113503774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,5120,0.13411911328633627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,10240,0.015087111128701104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,51200,0.13580000400543213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,768,0.016326222154829238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,51200,0.05488888753785027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,256,0.010616888602574667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,8192,0.03400711218516032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,256,0.005570666657553778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,5120,0.022142221530278523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,8192,0.022289777795473736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,1024,0.06839022371504042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,2560,0.010424888796276517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,7168,0.019498666127522785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,768,0.0041537777417235905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,6144,0.14305333296457926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,32,0.014656888114081489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,128,0.04277599851290385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,512,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,2048,0.04116977916823493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,3072,0.014518222875065274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,1536,0.010980444649855295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,2560,0.019262222780121695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,768,0.003726222034957674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,1024,0.15943911340501574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,512,0.00905955582857132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,5120,0.013725333743625216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,512,0.013615999784734515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,128,0.004060444318585926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,768,0.003768889026509391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,65536,0.06736089123619927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,256,0.006029333505365584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,32,0.0033555556502607134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,65536,2.821349249945747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,1536,0.07215555508931477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,2560,0.0094035557574696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,32,0.003232888877391815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,768,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,1024,0.004115555435419083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,32,0.029880887932247583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,2048,0.02446311049991184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,512,0.013448889056841532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,51200,0.08136977752049764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,10240,0.21751556131574842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,1024,0.1773617797427707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,2560,0.008207110895050896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,7168,0.0966915554470486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,128,0.005208000126812193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,5120,0.011208888557222156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,3584,0.00699733363257514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,2048,0.09767378038830227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,10240,0.01941066649225023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,8192,0.025991999440722998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,65536,0.13512889544169107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,3584,0.02885155545340644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,3584,0.008796444369686974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,16384,0.03554933269818624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,768,0.003666666646798452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,128,0.004584889031118817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,6144,0.024163555767801072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,2048,0.060250666406419545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,3072,0.10299911101659139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,1536,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,65536,0.12040089236365424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,65536,0.07456800010469225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,10240,0.12839555740356445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,65536,0.40204800499810117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,6144,0.07890933089786105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,512,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,16384,0.010045333041085137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,5120,0.011189333266682096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,2560,0.05765333440568712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,1536,0.06377244657940336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,4096,0.12976711326175266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,768,0.01442488862408532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,64,0.02003466586271922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,8192,0.04677155613899231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,3584,0.16915199491712782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,51200,0.39163822597927517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,1024,0.03536266750759549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,5120,0.3894853327009413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,7168,0.20750844478607178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,2048,0.014925332532988654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,3072,0.00794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,128,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,3584,0.03159733282195197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,32,0.00600355573826366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,3072,0.008077333370844523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,2048,0.01906488835811615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,768,0.02619733413060506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,256,0.02071199980047014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,8192,0.16814400090111625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,32,0.004902222090297275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,16384,0.12568267186482748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,32,0.004272888931963178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,4096,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,1536,0.01692800058258904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,8192,0.01479555500878228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,1024,0.03934489025010003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,8192,0.1391128831439548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,1536,0.10575733582178752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,768,0.00488711107108328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,2560,0.16642755932278103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,1024,0.025591111845440332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,3584,0.2094737821155124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,3072,0.04682044519318474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,256,0.00342399999499321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,1024,0.0058133333093590206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,3072,0.0281368891398112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,3072,0.009876444107956356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,1024,0.003990222182538775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,65536,0.8962382210625542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,4096,0.11043288972642686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,65536,0.09334666861428155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,1536,0.08646755748324925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,7168,0.640359984503852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,1024,0.005929777605666055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,3072,0.03615999884075589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,3584,0.04381510946485731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,32,0.017655111021465726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,12288,0.15173778269026014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,4096,0.008376888930797577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,7168,0.005873777800136142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,128,0.011172444456153445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,3072,0.08205688661999173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,6144,0.16712266869015166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,16384,0.0884408884578281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,4096,0.08161066638098823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,768,0.004970666848950916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,10240,0.02644888891114129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,128,0.010237333675225576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,2560,0.4847946696811252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,10240,0.4190631177690294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,6144,0.0068808889223469635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,65536,0.7998542255825467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,7168,0.011485333244005838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,3584,0.006739555547634761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,10240,1.4976106219821508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,1024,0.011071110765139261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,4096,0.04044177797105577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,6144,0.012395555774370829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,128,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,51200,7.694655524359809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,8192,0.49750577078925234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,64,0.00960888879166709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,10240,0.027640889088312786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,16384,0.23297243648105195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,16384,0.1632559961742825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,128,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,5120,0.03240888979699876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,512,0.009024000002278222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,3072,0.011013333168294696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,2048,1.2263164520263672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,5120,0.01107022249036365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,6144,0.014896000425020853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,6144,0.01716977854569753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,3072,0.026546667019526165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,12288,0.11135821872287327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,10240,0.4967173470391168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,3584,0.019910222954220243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,2560,0.048878222703933716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,10240,6.213008032904731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,12288,0.1273911131752862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,7168,0.03749422232309977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,768,0.0677493347062005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,51200,0.11441244019402398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,7168,0.015601777368121676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,32,0.005064888960785336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,64,0.0037635556525654266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,7168,0.02365511159102122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,2048,0.013717333475748697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,10240,0.06516089042027791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,16384,0.042048000627093844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,1536,0.01814399990770552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,12288,0.042915556165907115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,8192,0.01146044416560067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,1536,0.6442684597439235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,1536,0.0955137809117635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,51200,0.9352169036865234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,128,0.024098666177855596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,2048,0.033506668276256986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,3584,0.02489600082238515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,5120,0.012392889294359418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,5120,0.2645333343082004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,51200,0.20937955379486084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,4096,0.21141244305504692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,7168,0.1631351047092014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,3584,0.0739519993464152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,256,0.04792088932461209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,8192,0.08108088705274794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,10240,0.07579733265770806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,1024,0.007864889171388414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,2048,0.3488960001203749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,12288,0.04429955614937676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,2048,0.00565155554148886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,256,0.003818666769398583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,256,0.011382222175598145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,8192,0.02304711110062069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,64,0.002973333415057924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,5120,0.16527644793192545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,32,0.0028728888266616394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,6144,0.03771377934349908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,16384,0.03429244293106927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,10240,1.1618248621622722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,64,0.016328000360065036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,12288,0.010008888940016428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,512,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,6144,0.015455999308162265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,5120,0.022647110952271357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,5120,0.018575110369258456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,5120,0.008977777428097194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,8192,0.014946666028764514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,7168,0.030260443687438965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,51200,2.9304105970594616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,32,0.003068444422549672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,1536,0.09359111388524373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,1024,0.061940444840325244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,4096,0.024118221468395654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,4096,0.04861155483457777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,12288,3.698855929904514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,3584,0.1143875519434611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,8192,0.01496444476975335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,5120,0.11195733812120225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,1024,0.0069573331210348345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,2048,0.04068622324201796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,12288,0.024448888169394598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,768,0.05683466460969713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,12288,0.01831377711560991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,128,0.023248889380031165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,2048,0.015641777051819693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,65536,1.0306737687852647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,65536,1.4235732820298936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,5120,0.012357333468066322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,5120,0.0702800022231208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,1024,0.00599644457300504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,32,0.003129777808984121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,12288,0.15381956100463867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,3072,0.3400968975490994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,65536,0.1731955607732137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,768,0.18146577146318224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,6144,0.09488533602820502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,768,0.008388444781303406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,16384,0.06999288664923774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,4096,0.0058417779703934984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,2048,0.037414222955703735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,7168,0.18735022015041777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,4096,0.01957244508796268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,16384,0.13561511039733887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,128,0.004707555390066571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,3584,0.11715822749667698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,6144,0.00963644435008367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,2048,0.04067466656366984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,64,0.01960088809331258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,32,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,16384,0.008939555949634975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,12288,0.03040177623430888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,1536,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,7168,0.011021333436171213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,16384,0.45620086457994247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,2048,1.7188630633884008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,8192,0.4997688929239909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,12288,0.43491289350721574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,16384,0.22412443161010742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,1536,0.010384000009960597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,8192,1.2064168718126085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,51200,0.17568088902367485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,1024,0.018135999639829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,256,0.009354666703277165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,5120,0.15241155359480116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,256,0.07389155361387464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,8192,0.0088764445649253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,64,0.17139377858903673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,65536,3.9093093872070312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,256,0.0055440002017551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,256,0.007281777759393056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,6144,0.027427555786238775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,8192,0.03245688809288873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,1024,0.014135110709402295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,3072,2.3525502946641708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,12288,1.977834701538086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,32,0.015480889214409722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,256,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,256,0.0037475555307335327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,512,0.0169493340783649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,128,0.006948444578382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,128,0.005934222290913264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,3584,0.1629413366317749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,51200,0.6030106544494629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,32,0.014173333843549093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,1536,0.009003555609120263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,6144,0.028637333048714533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,5120,0.023017777336968318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,10240,0.044030222627851695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,128,0.009476444787449306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,16384,0.03050044510099623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,128,0.00544177782204416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,4096,0.02072622213098738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,512,0.010648888846238455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,512,0.04535555508401659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,64,0.002798222212327851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,8192,0.02550044490231408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,7168,0.005988444305128521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,51200,0.13684266143374973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,512,1.325625737508138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,16384,0.0224906669722663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,768,0.004085333396991094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,32,0.006807111203670502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,3584,0.008046222229798635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,512,0.046930667426851064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,512,0.009121778110663096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,256,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,12288,0.048750221729278564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,51200,0.1649564504623413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,128,0.005397333453098933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,64,0.0025617778301239014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,12288,0.01072266697883606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,512,0.0063582224150498705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,12288,0.03552888830502828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,4096,0.3092613220214844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,4096,0.2351484563615587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,256,0.030979557169808283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,512,0.010332444475756751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,8192,0.03817155626085069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,3072,0.009125333693292405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,7168,0.022127999199761286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,1024,0.01847999956872728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,3072,0.0081395556529363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,128,0.0075795559419526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,1024,0.0037253333462609183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,1024,0.0064488889442549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,6144,0.0181368887424469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,4096,0.009386666946940953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,1536,0.052397333913379245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,768,0.009520889156394536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,256,0.004503111044565837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,2560,0.05162577827771505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,3072,1.50219111972385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,128,0.003318222239613533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,5120,0.07728888591130574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,7168,0.006511999915043513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,8192,0.28478844960530597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,128,0.005818666683302985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,512,0.005865777946180767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,65536,0.7286666764153374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,512,0.05921510855356852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,10240,0.019398222366968792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,1536,0.005817777580685086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,768,0.003685333248641756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,3584,0.008022222254011366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,8192,0.006015111174848344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,768,0.01698755555682712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,65536,0.029824889368481104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,768,0.005254222287072076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,51200,0.2126337819629245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,768,0.03216088811556498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,1024,0.006028444402747684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,7168,0.19391022788153756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,3584,0.008017777568764156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,8192,0.012223111258612739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,6144,0.04029777646064758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,2048,0.0069351109365622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,128,0.003293333368168937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,2560,0.020276443825827706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,1024,0.011715555356608497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,2560,0.024847111768192712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,7168,0.014418666561444601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,6144,0.02496888902452257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,3584,0.07634044355816312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,4096,0.016172443826993305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,4096,0.024452444579866197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,10240,0.9062052832709419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,256,0.0062773335311147906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,65536,4.058658599853516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,5120,0.05508444375462002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,32,0.005888000130653381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,1024,0.032088001569112144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,3584,0.010130667024188572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,8192,0.05392711030112373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,1536,0.013991999957296582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,3584,0.0733777748213874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,128,0.010999110837777456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,10240,0.05852444304360283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,256,0.004399999976158142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,5120,0.21965600384606254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,1536,0.051832000414530434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,65536,0.23032710287306044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,2048,0.010819555156760745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,3072,2.5080738067626953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,32,0.0032257777121331957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,16384,0.3090924421946208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,65536,0.06680177979999118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,2048,0.00792622235086229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,4096,0.37494489881727433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,512,0.02460533380508423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,4096,0.7751226425170898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,5120,0.013667555318938361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,1024,0.003955555458863576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,7168,0.011048888994587792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,12288,0.021572444174024794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,4096,0.08532533380720352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,51200,1.502959993150499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,8192,0.035191999541388616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,128,0.01645599967903561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,8192,0.026707554856936138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,7168,2.544379552205404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,65536,0.4478115505642361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,128,0.012700444294346703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,4096,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,64,0.0027280000762807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,1024,0.010603555374675326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,2048,0.0358879996670617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,1536,0.030231999026404485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,2048,0.02146222194035848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,10240,0.2541297806633843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,3072,0.010562666588359408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,3584,0.22059999571906197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,1024,0.06422755453321669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,2560,0.008538666698667739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,128,0.022946667340066697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,768,0.022190221481853064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,10240,0.01776000029510922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,512,0.02167111138502757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,768,0.006809777683681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,10240,0.01776622235774994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,3072,0.019924443629052904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,65536,0.0804435544543796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,7168,0.10232710838317871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,512,0.009842666486899057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,10240,0.43336799409654403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,3584,0.203313774532742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,1024,0.11064533392588298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,2048,0.0188391109307607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,3584,0.039416889349619545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,1024,0.004303110970391167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,256,0.004765333400832282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,512,0.005731555736727185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,4096,0.05725510915120443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,3072,0.011876444849703046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,65536,0.0746213330162896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,768,0.0262773334980011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,3072,0.08265333043204413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,6144,0.3435386551751031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,512,0.015949333707491558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,5120,0.1096044447686937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,7168,0.11899022261301677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,512,0.004395555704832077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,3072,0.2208248906665378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,7168,1.4312701755099828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,256,0.003640888879696528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,1536,0.04320799973275927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,1024,0.020615999897321064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,10240,0.024903110331959192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,7168,0.4089297718471951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,65536,0.08531200223498875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,3072,0.023843554986847773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,128,0.022686221533351477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,768,0.26864979002210826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,7168,0.013039111263222165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,16384,0.2569626702202691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,6144,0.06571377648247613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,768,0.004271111140648524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,5120,0.022341334157519873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,3072,0.021478222476111516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,128,0.016711110870043438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,2048,0.008234666453467475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,1024,0.005842666659090254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,10240,0.016904888881577384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,768,0.015586665934986539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,2560,0.023747555083698694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,1024,0.040532443258497454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,768,0.00544177782204416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,12288,0.058950225512186684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,1024,0.016605334149466622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,51200,3.5916845533582897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,128,0.004941333499219683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,64,0.0714177754190233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,3584,0.007846222155623967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,8192,0.08236533403396606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,256,0.0029644444584846497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,256,0.012650666965378655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,10240,0.018836443622907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,3584,0.00831822223133511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,65536,1.6094293594360352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,51200,0.06886755757861666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,3072,0.039149334033330284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,51200,0.23802222145928276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,2048,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,8192,0.35418044196234805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,768,0.008435555630260045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,512,0.00813066628244188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,4096,0.10328177611033122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,8192,0.03130577670203315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,8192,0.028035554620954726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,12288,0.6305964257982042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,32,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,512,0.010978666444619497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,12288,0.008543999658690559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,7168,0.08279911014768812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,2560,0.02399377855989668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,2560,0.02358755634890662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,512,0.17030044396718344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,5120,0.025260445144441392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,6144,0.007295110987292395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,5120,0.012320000264379712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,512,0.007822222179836696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,6144,0.06607288784450956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,1536,0.010524444282054901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,5120,0.07034044133292304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,6144,1.8780081007215712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,65536,0.13558933469984266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,512,0.005245333330498801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,3072,0.012172444827026792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,3584,0.009307555854320526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,1024,0.006077333456940121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,256,0.016306666864289176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,16384,0.02535111043188307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,128,0.004624888714816836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,10240,0.059663110309176974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,512,0.011516444385051727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,51200,0.32712001270718044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,1024,0.023242667317390442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,1536,10.325247870551214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,64,0.012526222401195101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,6144,0.09699910879135132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,64,0.004001777619123459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,65536,1.4071138170030382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,10240,0.2798151175181071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,65536,0.7577395439147949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,4096,0.8233440187242297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,65536,0.5600639979044596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,768,0.0747306678030226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,65536,0.12656622462802464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,3584,0.06582844257354736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,65536,3.3651385837131076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,64,0.017430222696728177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,7168,0.3542533450656467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,1024,0.42417067951626247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,5120,0.08793066607581244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,5120,0.009565333525339762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,256,0.0047066667013698155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,2560,0.050204445918401085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,5120,0.011247999966144562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,512,0.006110222389300664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,32,0.009848888549539778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,768,0.07081333133909437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,128,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,1024,0.05891199906667074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,65536,2.9427119361029734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,1536,0.03749777873357137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,12288,0.023599111371570166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,2560,0.01742044422361586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,1024,0.11856355932023789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,256,0.020589333441522386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,128,0.003532444437344869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,10240,1.1286168628268771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,3584,0.03508799937036302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,8192,0.01051288925939136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,1024,0.010279110736317104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,16384,0.23569332228766549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,256,0.00944266633854972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,3072,0.02757511039574941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,6144,0.6895004378424751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,8192,0.033857779370413885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,51200,4.189494238959418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,51200,0.0851937797334459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,6144,0.019867555962668527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,3584,0.009407111340098912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,16384,0.06765511300828722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,128,0.010841777755154504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,256,0.005640888793600931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,4096,0.27407911088731557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,4096,0.02269511090384589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,8192,0.023034666975339253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,128,0.005939555664857228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,8192,0.012133333418104382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,6144,0.008534222013420528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,768,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,2560,0.029488000604841445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,51200,0.21121954917907715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,51200,0.13711022006140813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,10240,0.4951795472039117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,7168,0.035367111365000405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,32,0.015642666154437594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,16384,0.07301511367162068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,32,0.005051555732885997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,768,0.008741333252853816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,65536,0.27470845646328396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,3072,9.991437276204428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,2560,0.04931022061241997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,1024,0.007735999921957652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,5120,0.021730666359265644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,64,0.026429333620601233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,16384,0.07682755258348253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,64,0.0064088888466358185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,1024,0.00924711094962226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,7168,0.03749688797526889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,51200,0.7650533252292209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,16384,0.0222080002228419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,8192,0.014504889647165934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,3584,0.028774221738179524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,256,0.005538666827811136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,5120,1.0985067155626085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,2048,0.08953333563274807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,3072,0.024665777881940205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,8192,0.04529777831501431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,1536,0.008392888638708327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,64,0.0058373332851462895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,768,0.009914666414260864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,128,0.007724444071451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,2048,0.005967999911970562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,8192,0.03131377696990967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,2048,0.020415110720528495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,10240,0.5473795466952854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,1024,0.010825778047243753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,10240,0.451707575056288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,16384,0.022132444712850783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,7168,0.27706223063998753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,768,0.0038168889780839286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,12288,0.05617066886689928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,4096,0.013423110875818463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,16384,0.47626399993896484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,12288,0.032467554012934365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,2560,0.01460977726512485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,128,0.003068444422549672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,6144,0.01164177805185318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,5120,0.0069884442620807225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,768,0.02437244521247016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,10240,0.0639315578672621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,5120,0.014126222994592456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,6144,0.024283554818895128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,5120,0.0993280013402303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,51200,7.184366861979167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,6144,0.05431288811895582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,16384,0.6119093365139431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,8192,0.03340177734692892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,128,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,512,0.0036951113078329298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,512,0.005045333256324132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,16384,0.043390221065945096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,3072,0.006538666784763336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,16384,0.39631022347344297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,2048,0.00833422193924586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,2048,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,65536,0.0823226637310452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,6144,0.011192888849311404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,2048,0.04007110993067423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,3584,0.02330044408639272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,1536,0.013987556099891663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,768,0.00868444475862715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,3072,0.04671733246909248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,8192,1.5589947170681422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,12288,0.053314665953318276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,12288,3.0108587476942272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,128,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,12288,0.014044443766276041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,2048,0.008856000171767341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,768,0.011618666350841522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,2560,0.02180266711446974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,1536,0.030433777305814955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,12288,0.059207108285692006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,768,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,65536,0.10399999883439805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,4096,0.018135999639829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,4096,0.021734222769737244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,7168,0.012068444656001197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,4096,0.33034933937920463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,3584,0.2934604485829671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,16384,0.026491555902693007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,768,0.05433422327041626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,8192,0.21687377823723686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,12288,0.028554668029149372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,512,0.01865244408448537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,1536,0.008532444636027018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,2560,0.1779786613252428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,256,0.011033777561452655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,32,0.03219644559754266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,12288,0.9689902199639214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,65536,13.041015625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,51200,0.049490667051739164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,3072,0.042437334855397545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,3072,0.016540444559521146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,512,0.008729777402347988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,2560,0.014563555518786112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,32,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,32,0.13388800621032715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,12288,0.02750844425625271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,128,0.004882666799757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,65536,0.07178133063846164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,65536,4.361232757568359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,1536,0.030199110507965088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,6144,0.24963021278381348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,32,0.00685422205262714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,256,0.004317333300908406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,1536,0.013384000294738345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,7168,0.24347289403279623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,768,0.01584533353646596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,256,0.0070773329999711775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,2048,0.01348888874053955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,1024,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,8192,0.0398871103922526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,16384,1.4050782521565754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,10240,0.017975111802419026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,1536,0.010794666906197866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,256,0.024298666252030268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,512,0.005499555418888728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,5120,0.08625511328379314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,65536,0.13246488571166992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,3072,0.009568889107969072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,2048,0.009773333039548662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,10240,1.0342213312784831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,64,0.005032000028424793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,64,0.008015111088752747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,51200,0.8370604515075684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,256,0.17939644389682344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,3584,0.00978133330742518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,32,0.012093333734406365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,7168,0.1669457753499349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,128,0.0027466666781240036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,1024,0.025920889443821375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,51200,0.2401155630747477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,65536,0.45995198355780703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,7168,0.16325333383348253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,1536,0.012286221815480126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,51200,0.07986311117808025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,4096,1.9236053890652125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,4096,0.081386665503184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,4096,0.02030666669209798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,10240,1.5406986872355144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,6144,0.11426488558451335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,16384,0.01440355512830946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,51200,0.10907555288738674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,2560,0.10453689098358154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,65536,0.9953333536783854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,3584,0.008426666259765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,5120,2.2517609066433377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,2560,0.02477333280775282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,65536,0.08092711369196574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,32,0.04472088813781738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,32,0.0025191110455327565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,1536,0.007676444119877285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,32,0.0027466666781240036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,12288,0.8745182355244955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,1024,0.050204445918401085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,12288,0.02660888930161794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,51200,0.22677689128451875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,6144,0.04344977935155233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,768,0.003527111063400904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,32,0.005452444569932089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,2560,0.007198221981525421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,768,0.00423911131090588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,3072,1.1800720426771376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,10240,0.043487111727396645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,4096,0.01972888906796773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,2048,0.026168000366952684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,4096,0.14159022437201604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,2560,0.015250666273964776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,256,0.009106666677527958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,3072,0.6809840202331543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,768,0.004678222040335338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,1024,0.7056115468343099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,7168,0.04309422108862135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,1024,0.005803555664088991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,4096,0.02604177759753333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,3584,0.23348355293273926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,768,0.012903999951150684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,1024,0.011418666276666852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,64,0.01219733307758967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,1024,0.043861332866880626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,4096,0.011137777732478248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,1536,0.022495110829671223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,64,0.0027591110103660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,65536,0.31506310568915474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,8192,0.607934210035536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,51200,0.05232355660862393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,16384,0.24699287944369844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,6144,0.02121155626244015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,4096,0.03456444541613261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,512,0.003705777641799715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,51200,0.3377946747673883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,8192,0.13753867149353027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,5120,0.013944000005722046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,1024,0.02470933397610982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,65536,4.623271942138672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,4096,0.05056088831689623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,65536,0.16401155789693198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,16384,0.05534311135609945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,1024,0.05190933412975735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,64,0.008991111483838823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,65536,1.3934301800198023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,256,0.015023999743991427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,8192,0.019720888800091214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,3072,0.12760710716247559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,128,0.014161777165200977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,32,0.00700800038046307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,512,0.003798222376240624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,51200,0.05820000171661377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,8192,0.29330399301317006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,51200,0.3913066652086046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,768,0.003847111016511917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,768,0.06827377610736422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,12288,0.0698888897895813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,10240,0.023727110690540735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,7168,0.1969804498884413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,1536,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,768,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,7168,0.39785512288411456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,51200,0.3564106623331706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,1024,0.008541333178679148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,4096,0.02504088812404209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,12288,0.006786666810512543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,16384,0.046113779147466026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,256,0.005372444374693765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,32,0.006148444281684027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,1536,0.05893599987030029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,128,0.0029493332323100832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,5120,0.011287111375066968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,3072,0.41306045320298934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,5120,0.012044444680213928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,32,0.003444444388151169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,8192,0.0625822212960985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,256,0.015537778536478678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,1024,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,12288,0.1312862237294515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,3584,0.010030222435792288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,7168,0.07448622253206041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,10240,0.013532444834709167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,768,0.07093333535724215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,3072,0.5658444298638238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,8192,0.14744533432854548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,8192,0.26130665673149955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,7168,0.038487109873029925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,256,0.005210666606823604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,4096,0.0554328891966078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,6144,0.03477866782082452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,65536,0.21493777963850233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,256,0.010098666780524785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,12288,0.03450222147835626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,6144,4.294150034586589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,64,0.007479111353556315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,65536,1.239856931898329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,512,0.007229333122571309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,4096,0.08664355675379436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,10240,0.03491555651028951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,512,0.012721777790122561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,2560,0.028677334388097126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,768,0.01239377756913503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,7168,0.02918133470747206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,64,0.005511111269394557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,3584,0.5909111234876845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,4096,0.015455999308162265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,5120,1.7353839874267578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,51200,2.142001681857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,10240,0.19488267103830972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,6144,0.016351110405392118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,1024,0.006626666833957036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,512,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,32,0.1875173383288913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,256,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,10240,1.1075217988755968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,6144,0.015949333707491558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,2560,0.05538488758934868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,2048,0.05069066749678718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,16384,0.6229306856791178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,3584,0.16755199432373047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,6144,0.18886844317118326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,10240,0.013602666556835175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,6144,0.070378667778439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,7168,0.020376889242066275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,32,0.0035822221802340615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,1536,0.008656889200210571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,2048,0.020741333564122517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,65536,1.742529763115777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,3072,0.009341333475377824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,4096,0.012126222252845764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,8192,0.13693866464826795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,768,0.004592000196377437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,128,0.004473777694834603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,51200,0.15958133008744982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,51200,0.057019558217790395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,5120,0.029189334975348577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,2560,0.012493333054913415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,4096,0.008295111358165741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,1536,0.01408266690042284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,128,0.011411555939250521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,2560,0.007301333049933116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,7168,0.021969777014520433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,2048,0.02389066749148899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,512,0.004632888982693355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,10240,2.0856196085611978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,2560,0.01588800052801768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,768,0.007263110743628607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,512,0.04938844508594937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,512,0.0038951109680864546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,32,0.002846222163902389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,4096,0.5635866589016384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,256,0.06045866674847073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,7168,0.008079999850855933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,3584,0.013139555851618448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,512,0.016193777322769165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,12288,0.017376888129446242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,1024,0.004044444610675176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,7168,0.08520888619952732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,7168,0.01679999960793389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,2048,0.02460266649723053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,6144,0.02463022205564711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,3584,0.03244355652067397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,51200,2.440247005886502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,64,0.0094373333785269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,3584,0.1073839995596144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,10240,0.027053332991070215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,7168,0.04714399907324049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,32,0.005551110953092575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,1024,0.006512889017661412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,16384,0.35643911361694336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,6144,0.012848888834317526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,4096,0.008283555507659912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,768,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,65536,0.026110221942265827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,32,0.002571555475393931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,4096,0.03344710999064975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,7168,0.10395644770728217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,128,0.009671111073758867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,51200,0.14725511603885227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,512,0.013844444519943662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,3072,0.5689457787407769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,8192,0.04867022236188253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,3072,15.82783677842882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,16384,0.024188444018363953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,5120,1.6913021935356989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,51200,0.6454337967766656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,1536,0.036517331997553505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,2048,0.007111999723646376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,4096,0.021998221675554912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,51200,0.036947555012173124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,10240,0.02013066742155287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,12288,0.07561600208282471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,7168,2.4253209431966147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,10240,0.06081777811050415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,2560,0.010888888604111142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,1024,0.19134933418697783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,1024,0.05217866765128242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,51200,0.09167644712660049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,8192,0.523904906378852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,64,0.0030053332448005676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,3072,0.00923911150958803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,7168,0.03802311089303758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,128,0.026479111777411565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,5120,0.27243023448520237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,128,0.02460799945725335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,256,0.005616888817813661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,3072,0.010094222095277574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,65536,2.4061465793185763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,3072,0.008234666453467475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,3584,0.013731555806265937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,3072,0.06061422162585788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,1024,0.0047600000268883174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,4096,0.24825776947869194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,65536,20.963470458984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,12288,0.1888328923119439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,7168,0.1016462246576945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,4096,0.5323528713650173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,1536,0.02164266672399309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,10240,0.047621332936816745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,8192,0.011613333390818702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,65536,0.9168933232625326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,16384,0.5502755377027724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,51200,0.3088737858666314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,256,0.007978666987684038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,3584,0.007870222131411234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,1024,0.005737777799367905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,2560,0.030936890178256567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,12288,0.044906665881474815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,12288,0.07340533203548856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,1536,0.012804444465372296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,768,0.005324444423119227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,51200,0.27038576867845326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,1024,0.020953777763578627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,256,0.008767999708652496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,128,0.0028035555862718155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,512,0.007983999947706858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,12288,0.017982222967677645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,768,0.05456533365779453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,256,0.013251555462678274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,768,0.03798310955365499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,2560,0.0712604456477695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,6144,0.3881848918067084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,6144,0.041391998529434204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,3072,0.008082666330867344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,10240,0.016324443949593436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,16384,0.10251200199127197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,3072,0.20265511671702066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,51200,0.41696622636583114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,12288,0.02751022246148851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,65536,0.26302043596903485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,768,0.0347635547320048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,65536,0.2141795555750529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,6144,0.016302223006884258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,3584,0.012201777762836881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,8192,0.0732426643371582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,65536,0.14805511633555093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,32,0.006063111126422882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,1024,0.300014230940077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,6144,0.026318222284317017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,3584,0.3652391168806288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,5120,0.0499875545501709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,51200,0.37021954854329425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,12288,0.03840622305870056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,51200,1.865951961941189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,7168,0.015146666102939181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,2048,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,128,0.014627556006113688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,12288,0.3971795505947537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,64,0.013559111290507846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,32,0.0033733333564466904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,16384,0.03397599856058756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,3072,0.05256355471081204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,1536,0.052184889713923134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,1536,0.37024445003933376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,64,0.046987556748920016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,6144,0.15603378083970812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,1536,0.07586489121119182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,2048,0.28106755680508083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,5120,0.04173600011401706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,768,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,3584,0.05442755421002706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,7168,0.014418666561444601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,32,0.004086222085687849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,5120,0.14703288343217638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,256,0.005388444496525659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,2048,0.016588444511095684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,32,0.0028231110837724474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,6144,0.02314311100376977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,3072,0.00756177802880605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,8192,0.06641511122385661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,51200,0.03737866547372606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,128,0.0052266667286554975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,128,0.0038968887594011095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,51200,0.06350933180914985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,64,0.13770933945973715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,12288,3.6144896613227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,128,0.0313351121213701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,2560,0.012626666989591388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,256,0.00980799976322386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,8192,0.025062221619817946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,1024,0.026390223039521113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,8192,0.008206222620275285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,256,0.0033555556502607134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,1536,0.26987377802530926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,768,0.5600346459282769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,7168,0.034244444635179304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,8192,0.04531200064553154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,6144,0.009120889008045197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,1536,0.09655733240975274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,768,0.05221155616972181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,1536,0.006206222292449739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,16384,0.9539688958062066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,64,0.017816000514560275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,768,0.012103999654452005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,1536,1.307435565524631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,3072,0.03182222114668952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,256,0.005422222117582957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,4096,0.20859821637471518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,32,0.012117333710193634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,1024,0.024099555280473497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,10240,0.05527111225657993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,51200,0.059193776713477246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,3584,0.04974400003751119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,51200,0.28368621402316624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,5120,1.5634987089369032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,12288,0.02826311190923055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,3584,0.023079110516442194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,2048,0.007834666305118138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,7168,0.04860533277193705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,51200,0.13634399573008218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,10240,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,10240,0.008428444465001425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,6144,0.008586666650242275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,6144,0.0965333316061232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,1536,0.01625333395269182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,3584,0.017674666312005784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,8192,0.03214933474858602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,1024,0.012708444562223224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,128,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,512,0.026239999466472205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,768,0.009940444595283931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,1536,0.037419554260041975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,5120,0.28299911816914874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,5120,0.10415911012225681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,512,0.0068897778789202375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,512,0.015390222271283468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,7168,0.019783111082182992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,51200,1.850830290052626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,2048,0.02329066726896498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,7168,0.06611466407775879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,7168,0.26939911312527126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,1536,0.010264000131024255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,1536,0.006666666517655055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,2048,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,51200,1.8522781795925563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,8192,0.21089422702789307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,4096,0.027151111099455092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,51200,0.3328568935394287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,51200,2.306380377875434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,768,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,6144,30.126641167534725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,2560,0.035290665096706815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,65536,0.907416025797526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,65536,0.3318551116519504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,7168,0.11407200495402019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,1536,0.006268444574541516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,4096,0.07413866784837511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,2560,0.01772622267405192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,768,0.007739555504586961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,2048,0.07530577977498372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,16384,0.5363235473632812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,8192,0.03630044394069248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,64,0.050980445411470204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,32,0.0042364444169733255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,32,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,2048,0.03284800052642822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,32,0.003212444484233856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,2560,0.009134222235944536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,10240,0.05482044484880236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,128,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,8192,0.023327999644809302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,16384,0.14980710877312556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,1024,0.02530488868554433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,128,0.004291555533806483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,6144,0.04221955604023404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,256,0.01552977826860216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,16384,0.013939556148317126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,128,0.00742133292886946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,5120,0.01699733402993944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,256,0.01128266668981976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,5120,0.01000088867213991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,7168,0.03339022066858079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,1024,0.004848888764778773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,51200,0.5518710878160265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,768,0.0052915554907586836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,1024,0.009212444225947062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,10240,0.047319110896852284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,6144,0.026363554928037856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,4096,0.02529688841766781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,3072,0.01573155489232805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,4096,0.00849333322710461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,6144,0.01629688839117686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,2560,0.019467555814319186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,1024,0.010647999743620554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,4096,0.01482222146458096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,2048,0.02313866714636485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,51200,0.4419875674777561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,3584,0.1708231104744805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,128,0.042993777328067355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,2048,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,16384,0.031268444326188825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,512,0.0057653333577844836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,4096,0.20371199978722465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,2048,0.04748444424735176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,3584,0.6022071308559841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,5120,0.2404222223493788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,6144,0.01958933307064904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,32,0.0025555555605226094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,10240,0.03329955538113912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,1536,0.1541528834236993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,1024,0.008371555142932469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,512,0.00518222235971027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,512,0.003846222327815162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,3072,0.007997333175606197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,1536,0.012552888856993781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,51200,0.7930151091681586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,16384,0.11593243810865615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,65536,0.24876266055636934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,1536,0.08437155352698432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,7168,0.011384000380833944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,2048,0.009930666950013902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,2560,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,128,0.004604444321658877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,4096,0.20498577753702799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,65536,5.721728006998698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,256,0.015100444356600443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,6144,0.01350400017367469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,4096,0.01591822173860338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,256,0.0032346666687064697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,128,0.03518844313091702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,256,0.0031191110610961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,512,0.0032248888164758682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,2560,0.04000533289379544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,10240,0.014004444082578024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,768,0.004546666724814309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,768,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,7168,0.02903111113442315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,3584,0.027259555127885606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,12288,0.04064977831310696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,256,0.0053680001033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,51200,0.6661778026156956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,64,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,4096,0.026689777771631878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,2560,0.045703109767701894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,512,0.04692799846331278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,6144,0.023174222972657945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,64,0.0026400000270869997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,256,0.008789333204428354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,65536,11.485945807562935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,10240,0.46904887093438047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,32,0.008764444126023186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,512,0.006896889044178857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,65536,0.36211644278632266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,10240,0.4603457980685764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,10240,0.07465955283906725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,2560,0.006542222367392645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,8192,0.02160355614291297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,10240,0.04003377755482992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,2048,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,8192,0.3520035478803847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,51200,0.599812454647488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,64,0.005512889060709212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,64,0.0030257778449190986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,1536,0.0181057784292433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,51200,1.4200240241156683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,32,0.00563733321097162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,768,0.038294222619798444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,7168,0.014447111222479077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,4096,0.028560890091790095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,128,0.13039199511210123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,6144,2.1250347561306424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,256,0.0038151111867692736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,65536,1.1374008390638564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,1536,0.017224889662530687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,10240,0.03019022279315525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,4096,0.042790220843421094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,2560,0.023772444989946153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,12288,0.7835689120822482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,128,0.008533333738644918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,2560,0.008906666603353288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,4096,0.03832799858517117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,2560,0.04759644468625387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,2560,0.010412444670995077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,128,0.026535999443795946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,2560,0.01187999960449007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,4096,0.009815110928482479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,2560,0.051945779058668345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,32,0.007520889242490132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,12288,0.05242844422658285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,7168,0.007213333414660559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,256,0.021143111917707656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,65536,0.8659271134270562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,65536,0.6211502287122939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,32,0.032405333386527166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,51200,0.0856871141327752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,3072,0.013479111095269522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,5120,0.031110223796632554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,12288,0.040186666780047946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,3584,0.02260355485810174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,65536,0.18948533799913195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,512,0.0033697777738173804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,12288,1.401277330186632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,32,0.00945688866906696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,16384,0.010976889067225985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,12288,0.24924445152282715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,1024,0.023727999793158636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,512,0.008170666793982187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,8192,0.014558222558763293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,7168,0.02230311102337307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,1536,0.03278844555219015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,128,0.005120000077618493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,16384,0.24319820933871797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,256,0.007223111059930589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,3072,0.006597333484225803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,64,0.0026595555245876312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,3584,0.01800622211562263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,1024,0.015591111448076038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,3072,0.018188445104493033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,65536,0.07012089093526204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,7168,0.09456355704201592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,6144,0.022436444958051045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,768,0.06095288859473335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,2560,0.09067289034525554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,65536,0.14727289146847194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,8192,0.06026666694217258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,3584,0.1388053364223904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,4096,0.11920000447167291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,10240,0.048230220874150596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,16384,0.034953776333067156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,2048,0.024142222272025213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,6144,0.01776266594727834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,12288,0.019175110591782462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,2048,0.005264889034960005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,4096,1.375147607591417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,1536,0.2542275587717692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,16384,0.032835556401146784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,256,0.4547102186414931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,3584,1.470492468939887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,1024,0.007812444534566667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,1536,0.5447102122836643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,3072,0.02616888946957058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,6144,0.05469510952631632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,768,0.006759111252095964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,3072,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,51200,0.9330231348673502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,2560,0.010817777779367236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,6144,0.023943111300468445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,512,0.007963555554548899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,16384,0.29348799917432994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,768,0.027758222487237718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,2048,0.006302222195598815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,10240,0.145325329568651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,1024,0.014461333552996317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,256,0.009197333620654212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,512,0.04923289020856222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,5120,0.03481600019666884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,65536,0.19447022014194063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,2048,0.006761777732107375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,1536,0.01723377737734053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,16384,0.241979546017117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,256,0.0031928889867332247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,5120,0.01070666644308302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,51200,0.016366221838527255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,51200,0.06601511107550727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,512,0.2924355665842692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,32,0.005131555514203178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,51200,0.053048001395331494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,8192,0.1569262213177151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,12288,0.012042666474978128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,7168,0.40030399958292645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,5120,0.014483556151390076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,51200,0.0918017758263482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,64,0.0026008888251251644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,3072,0.013162666724787818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,3072,0.006731555693679386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,16384,0.02276177704334259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,32,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,12288,0.021631111701329548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,12288,0.0835262205865648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,128,0.1592080063290066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,32,0.002611555573013094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,2560,0.05275377631187439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,2560,0.08114488919576009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,4096,0.013995556367768182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,512,0.011167999770906238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,7168,0.008222222328186035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,128,0.01073688848151101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,7168,0.011864888999197217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,512,0.00720711135201984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,4096,0.03223199976815118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,51200,1.719127125210232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,1024,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,16384,0.03476266728507148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,768,0.023211555348502264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,512,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,128,0.0034844444857703317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,7168,0.013771555489963956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,12288,0.6518701977199978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,128,0.0028373334142896864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,8192,0.026034666432274714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,12288,0.11180267069074844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,256,0.015927111109097798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,65536,3.392907460530599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,51200,0.08166044288211398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,64,0.006576889091067844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,4096,0.15032977528042263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,3072,0.015632000234391954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,6144,0.01815555493036906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,32,0.006264888991912206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,768,0.036941332949532404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,1024,0.12760978274875218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,256,0.02438755498992072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,256,0.002842666581273079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,1024,0.04999644557634989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,256,0.014941333068741692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,2048,0.007535999847782983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,65536,0.07797155777613322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,8192,1.0216471354166667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,8192,0.3010835647583008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,4096,0.0063777777055899305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,65536,0.06908533308241102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,16384,0.6207697656419542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,5120,0.16448622279696995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,6144,0.4216586748758952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,16384,0.1303688950008816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,512,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,512,0.022410665949185688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,7168,0.006411555740568373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,32,0.005392000079154968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,256,0.00332266671790017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,4096,0.03587466809484694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,1024,0.01127377814716763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,1024,0.007937777373525832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,6144,0.02644266684850057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,512,0.005156444592608346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,8192,14.123102823893229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,2560,0.1384399996863471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,8192,3.5824063618977866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,8192,0.1088177760442098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,1024,0.028762665059831407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,5120,0.032165331972969904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,6144,0.05273955398135715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,65536,1.2462569342719185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,2048,0.24228533109029135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,256,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,5120,0.029648890097935993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,51200,0.1194231112798055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,2560,0.16803288459777832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,65536,0.1690817806455824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,6144,0.009763555394278632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,10240,0.06624088684717815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,8192,0.23251554701063368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,256,0.0046773333516385825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,1024,0.016736888223224215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,4096,0.008760888543393876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,1536,0.011687111523416309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,2048,0.017236444685194228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,16384,0.018231110440360177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,4096,0.008302222523424361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,512,0.05067288875579834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,5120,0.037119110425313316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,51200,0.057349330849117704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,51200,1.3771946165296767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,64,0.02790044413672553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,16384,0.04632622334692213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,10240,2.278752009073893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,64,0.005494222044944763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,2048,0.01292711082432005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,1024,0.05631733602947659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,768,0.0081404447555542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,32,0.0036551112102137674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,7168,0.08191821972529094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,16384,0.2779208819071452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,128,0.0027226667023367355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,32,0.0028782222006056043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,6144,0.04989155464702182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,2560,0.14761422740088567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,16384,2.4527946048312717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,8192,0.022456889351209004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,10240,0.020999999509917364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,3072,0.020602666669421725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,65536,0.3257955445183648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,768,0.010435555544164447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,64,0.003007111036115222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,1536,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,8192,0.03348799877696567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,12288,0.037667555941475764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,64,0.002603555512097147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,12288,0.0959191123644511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,1536,0.004620444443490771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,6144,0.021819555097156104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,5120,0.022216889593336318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,1536,0.5752026769849989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,128,0.004965333475006951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,7168,0.2691902319590251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,256,0.005774222314357758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,3072,0.018783110711309645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,8192,0.014331556028789945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,1536,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,64,0.005418666534953647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,4096,0.011249778171380361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,1024,0.010498666928874122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,3584,0.06171377499898275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,8192,0.3695128758748372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,12288,0.007651555869314406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,32,0.005943999936183293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,2048,0.009791110952695211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,128,0.00793688909875022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,1536,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,2048,0.10203999943203396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,16384,0.17473422156439888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,512,0.007902222375075022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,51200,2.5158142513699002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,16384,0.03343910972277323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,51200,0.16245333353678384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,128,0.007667555577225155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,3072,0.008519111408127679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,512,2.1943787468804254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,7168,0.053357332944869995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,32,0.0090506664580769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,4096,0.04976355698373583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,512,0.06859911150402494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,12288,0.2061848905351427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,256,0.008002666963471307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,2560,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,2560,0.00944800012641483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,256,0.3056844340430366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,8192,0.918940438164605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,6144,0.02510044475396474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,8192,0.16437778207990858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,7168,0.08839911222457886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,8192,0.01944977707333035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,512,0.008759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,64,0.0034968890249729156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,3584,0.030360887447992962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,256,0.04583999845716688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,1024,0.08248444398244222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,51200,0.32937600877549916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,3584,0.01609866652223799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,3072,0.06892533434761895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,128,0.014204444156752693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,8192,0.06317333380381267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,2048,0.006032889087994893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,12288,0.254351986779107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,4096,0.01757244434621599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,12288,0.35159820980495876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,12288,0.0860862202114529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,3584,0.010043555663691627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,256,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,3584,0.008262222011884054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,768,0.013554666605260638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,256,0.0047457776963710785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,12288,0.11692533228132461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,8192,0.03523022267553542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,3584,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,10240,0.015249777171346875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,1536,0.029628445704778034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,4096,0.019886222150590684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,10240,0.019693333241674636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,65536,1.7293404473198786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,16384,0.09028177791171604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,8192,0.5599306424458822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,3072,0.006119110931952794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,5120,0.23976267708672416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,51200,0.1893413331773546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,768,0.006625777731339137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,51200,0.023727110690540735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,3072,0.14321245087517634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,3072,0.07790933052698772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,3072,0.014136888914638095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,1024,0.00424799985355801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,128,0.0041928887367248535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,3072,0.10694044166141087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,128,0.0027297778675953546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,2560,0.012601777911186218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,12288,0.07102666960822211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,7168,0.023598222268952265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,2560,0.07451911105049981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,51200,0.2411937713623047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,128,0.007292444507280986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,2560,0.01830400029818217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,2048,0.014927110738224454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,128,0.005407111098368962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,6144,0.02039377722475264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,65536,0.334020455678304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,128,0.0074755557709270054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,1024,0.15321955415937635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,16384,0.03750133183267381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,4096,0.018762666318151686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,64,0.0027093332674768236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,128,0.05743022097481621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,3072,1.2029511133829753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,65536,0.05519555674658882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,51200,31.477884928385418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,1536,0.008423111504978603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,768,0.005953777581453323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,10240,0.09584533505969578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,6144,0.01955911186006334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,8192,0.10145066844092475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,16384,0.062182221147749156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,5120,0.008143111235565608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,512,0.0041164445380369825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,768,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,7168,0.03067644437154134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,10240,0.029208001163270738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,3072,0.02038666605949402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,65536,0.1877511077457004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,128,0.005152889009979037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,1536,0.08846844567192926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,10240,0.13812355200449625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,16384,0.025335111551814612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,7168,0.0064035554726918536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,3584,0.007344889144102733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,2560,0.021673777037196692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,32,0.0026942222482628296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,512,0.005976000179847081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,1024,0.09024266401926677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,1536,0.013605333036846586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,6144,0.02082577844460805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,3072,0.0075502221783002215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,5120,0.01589866644806332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,1024,0.006872889068391588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,16384,0.1902302238676283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,65536,0.02944177720281813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,3584,0.11450844340854222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,10240,0.10292355881796943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,32,0.0027057776848475137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,512,0.004928888960017098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,7168,2.1456987592909074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,2560,0.1672604481379191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,3072,0.04997511042488945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,3584,0.022277333670192297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,5120,0.026408889227443274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,1536,0.04180977741877238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,2048,0.027348443865776062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,8192,0.33185333675808376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,65536,1.7363182703653972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,512,0.004008888784382078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,6144,0.510857794019911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,128,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,4096,0.011372444530328115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,7168,0.012492444780137805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,1024,0.049397332800759204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,2048,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,512,0.03398133317629496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,768,0.0340008901225196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,1024,0.011718221836619906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,51200,0.06829422050052218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,2560,0.007092444433106317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,8192,0.206749333275689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,768,0.005474666754404704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,2048,0.006842666616042455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,65536,0.06827911403444079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,5120,0.018166222506099276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,8192,0.02390755547417535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,12288,0.11087644762463039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,8192,0.024413333998786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,51200,0.1717475520239936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,512,0.004280000097221798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,768,0.01348444405529234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,1024,0.008743999732865227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,6144,0.24000443352593315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,7168,0.03214755654335022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,12288,6.259373135036892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,3584,0.16232444180382624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,256,0.0029688889367712867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,2560,0.10761955711576675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,8192,0.08778399891323513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,6144,0.018231999542978074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,8192,0.31719644864400226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,7168,0.1382853322558933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,3584,0.03177155719863044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,1536,0.04334933227962918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,2048,0.010935111178292168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,2048,0.07546755340364245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,1536,0.010355555348926121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,65536,23.869664510091145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,768,0.007339555356237624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,1536,0.09530844291051228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,16384,0.02957688768704732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,8192,0.0660159985224406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,512,0.003488000068399641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,51200,7.737689548068577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,10240,0.1091502242618137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,7168,0.26322489314609104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,128,0.0047662220895290375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,8192,0.014507555299335055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,2048,0.013456000222100152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,10240,0.021581333544519212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,512,0.010018666585286459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,16384,11.017116970486112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,768,0.3182408809661865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,6144,0.017136888371573556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,512,0.008719111482302347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,4096,0.010811555716726514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,3584,0.0858115553855896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,256,0.22025067276424834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,2048,0.006144888699054718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,512,0.007816000117195977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,512,0.016673778494199116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,2048,0.007703999678293864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,768,0.013405332962671915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,65536,18.4329833984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,1536,0.03706666827201843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,16384,2.8385840521918406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,16384,0.033193777004877724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,1536,0.2991822295718723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,12288,0.017656000124083627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,1024,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,16384,0.08195555872387357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,2560,0.008993777963850234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,512,0.025892444782786902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,10240,0.05104355679617988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,51200,16.45389133029514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,768,0.05379466546906365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,65536,0.1474835607740614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,7168,0.11365244123670791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,3072,0.04823644293679131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,512,0.25057511859469944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,768,0.004355555607212914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,12288,4.488458845350477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,8192,0.10308533244662815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,10240,0.05072177780999077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,2048,0.006224000205596288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,1536,0.01309333327743742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,7168,0.06699911091062757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,6144,0.014076444837782117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,1024,0.017312000195185345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,7168,0.03536622060669793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,10240,0.015468445089128284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,16384,0.11124355263180202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,512,0.003436444534195794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,5120,0.019235556324323017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,3584,0.007884444461928474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,128,0.006908444480763541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,3072,0.091575112607744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,256,0.005528888768619961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,51200,3.047483656141493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,16384,0.0820195542441474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,51200,0.19462222523159453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,256,0.0033111110743549135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,32,0.0030177777840031516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,8192,0.022252443763944838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,1024,0.01847199930085076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,1536,0.009513777991135916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,4096,0.019691555036438834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,7168,0.030806223551432293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,4096,0.03019022279315525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,3072,0.05835644404093424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,10240,0.11563200420803493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,32,0.02347822156217363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,2560,0.008183110919263627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,512,0.013854222165213691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,4096,0.008404444489214156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,51200,2.142637252807617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,2048,0.0107004443804423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,5120,0.022210665875011023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,12288,0.028718223174413044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,2048,0.6109671062893337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,64,0.010400888820489248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,512,0.0033031110134389666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,768,0.009184000392754873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,3072,0.00867911097076204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,512,0.0418177776866489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,1536,0.004975999808973736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,7168,0.46584177017211914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,32,0.0123831108212471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,10240,0.05325066712167528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,512,0.0034008889148632684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,256,0.0063777777055899305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,65536,0.08395555284288193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,512,0.0035617777870761026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,5120,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,2048,0.02068711154990726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,2560,0.02308622168170081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,10240,0.018276444739765592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,32,0.013604444762070974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,256,0.004345777961942884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,6144,0.020662221643659804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,10240,0.007061333292060428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,2560,0.04853333367241753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,51200,0.5943902333577474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,65536,0.1702115535736084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,51200,0.0644657810529073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,3072,0.12153777811262344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,3584,0.029496000872717962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,32,0.002704888789190186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,3584,0.28433865971035427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,512,0.008738666772842407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,7168,0.7447919845581055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,7168,0.025115556187099878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,4096,0.1411679983139038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,768,0.005533333453867171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,512,0.0034293333689371743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,32,0.01404888927936554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,2560,0.024422221713595923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,512,0.024271999796231587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,7168,0.04765511221355862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,64,0.0034479999707804788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,10240,0.042528887589772545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,7168,0.06457333432303534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,6144,0.019202666150199044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,1536,0.007553777760929531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,4096,0.006367110957702001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,8192,0.028076443407270644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,51200,8.85658942328559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,5120,0.1746000051498413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,8192,0.4402933385637071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,512,0.006467555546098285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,1536,0.052892443206575185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,2560,0.016152888536453247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,51200,0.08291822009616427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,65536,0.49022754033406574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,64,0.13172800011105007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,51200,0.059954669740464955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,512,0.0037235555549462638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,768,0.004967111266321606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,2048,0.022970666488011677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,3072,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,3584,0.09478755791982015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,12288,0.032122668292787336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,128,0.005103999955786599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,12288,0.0399982233842214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,256,0.0048062221871482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,768,0.006660444455014334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,3584,0.06951733430226643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,65536,2.6595804426405167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,10240,0.020336889558368258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,5120,0.009073778159088558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,1024,0.007380444142553542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,2560,0.008369777765538957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,3072,0.06548355685340033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,5120,0.018034666776657104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,4096,0.04487999942567614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,1024,0.031054221921496924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,128,0.011156444748242697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,3072,0.06486666864818998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,5120,0.01094133324093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,1024,0.06588888830608793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,8192,0.014525334040323893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,3072,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,8192,0.0064693333374129395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,32,0.003939555750952827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,8192,0.07557777563730876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,7168,0.020231111182106864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,65536,0.4095591174231635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,7168,0.0372977786593967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,3072,0.10264711247550116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,7168,0.21221066845787895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,3072,0.052778667873806424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,2048,0.14460178216298422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,3072,0.015593777100245157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,1024,0.09948177470101251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,6144,0.020587555236286588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,256,0.003859555555714501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,768,0.013351110948456658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,1024,0.005946666830115848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,128,0.02367111047108968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,2048,0.05684000253677368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,7168,0.04504088891877068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,51200,0.31445243623521596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,16384,0.08932533529069688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,16384,2.028079138861762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,8192,0.26005956861707896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,8192,0.012056888805495368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,16384,0.352816899617513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,64,0.006666666517655055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,16384,0.11296978261735703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,5120,0.40744000011020237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,12288,0.010581333604123859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,2560,0.04548977812131246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,10240,0.3022186756134033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,65536,0.15920978122287327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,7168,2.344890594482422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,65536,2.4355057610405817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,10240,0.011536888778209686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,2048,0.00885333369175593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,1536,0.07194133599599202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,4096,0.038020445240868464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,16384,0.6243164274427626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,128,0.0036675557494163513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,2560,0.007255111303594377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,10240,0.03243911266326904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,768,0.027289777994155884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,8192,0.025756445195939805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,512,0.14627822240193686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,1536,0.02648799949222141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,512,0.011936888926559024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,51200,0.05579466952217949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,10240,0.47522931628757054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,6144,0.014581332604090372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,6144,0.014474666780895658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,4096,0.02610933283964793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,12288,0.5689715809292263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,16384,5.785758972167969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,1536,0.034280889564090304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,1024,0.025254221426116094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,12288,0.020611556039916146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,16384,0.03886222177081638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,512,0.00675022229552269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,2560,0.009105777574910058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,1536,0.08784266975190903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,12288,0.007214222517278459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,12288,0.01887911061445872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,768,0.0039057777159743835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,1536,0.005606222069925732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,3072,0.08429955773883396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,65536,0.7180328898959689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,2560,0.20062843958536783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,7168,5.845920986599392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,10240,0.22178844610850015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,32,0.0069955554273393415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,8192,0.017104888955752056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,256,0.007668444679843054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,7168,0.022328888376553852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,3072,0.01309866706530253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,51200,1.8591066996256511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,512,0.008061333662933773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,12288,0.31806932555304634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,256,0.004921777794758479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,10240,0.9505137337578667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,1024,0.15191645092434353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,7168,0.024526221884621516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,10240,0.018464000688658822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,256,0.008572444319725037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,16384,0.3668915430704753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,10240,0.021723555194007024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,256,0.010108444425794814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,5120,0.018811555372344125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,3072,0.02980977627966139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,10240,0.14224889543321398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,2560,0.01885777711868286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,1024,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,51200,0.10905422104729547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,10240,0.13360355959998235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,16384,0.02844711144765218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,51200,0.9616657892862955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,8192,0.04783911175198025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,4096,0.0766177773475647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,128,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,65536,0.1761528915829129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,5120,0.14884177843729654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,512,0.004504000147183736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,7168,0.10238044129477607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,16384,0.13677777184380427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,512,0.010942222343550788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,512,0.014176888598336114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,16384,0.011607999602953592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,7168,0.01166488892502255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,256,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,3584,0.006778666542636023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,3584,0.05081511206097073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,2560,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,5120,0.024271999796231587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,1536,0.008348444269763099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,2048,0.03449066811137729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,32,0.1341022253036499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,10240,0.13795377148522273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,2048,0.012207110722859701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,512,0.0031715554909573663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,1024,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,2048,0.08018044630686443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,65536,0.14663910865783691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,51200,0.0835075577100118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,10240,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,1024,0.0446631113688151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,256,0.006339555399285422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,10240,0.44138489829169375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,4096,0.12349155214097764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,1024,0.005355555564165115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,12288,4.2278158399793835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,1536,0.04905866583188375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,512,0.01756977703836229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,3072,0.012431110772821637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,256,0.0053004444473319584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,3584,0.08243199851777819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,128,0.0027751111321979095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,4096,0.04387288954522875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,32,0.008535111116038429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,6144,4.618201785617404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,256,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,7168,0.017743999759356182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,768,0.0041306668685542215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,7168,0.024455111887719896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,2048,0.013175110850069257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,2560,0.0864844454659356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,2048,0.0068986668354935115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,256,0.013074666261672974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,51200,0.2794240050845676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,64,0.007442666424645319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,1024,0.0040960001448790235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,5120,0.014211555322011312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,3072,0.07879555225372314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,768,0.0057884446448749965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,768,1.20377688937717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,8192,0.4175395435757107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,7168,0.01332444449265798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,128,0.005355555564165115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,3584,0.18435111310746935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,768,0.015832889411184523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,1024,0.010399999717871347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,12288,0.038740442858801946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,10240,4.160150316026476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,2560,0.011791110866599612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,128,0.013360888593726687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,512,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,768,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,65536,0.014542222023010254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,4096,0.06172177526685926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,8192,0.014151111245155334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,1024,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,51200,0.13235289520687527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,1024,0.01256444470749961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,3584,0.020089776979552377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,768,0.05893333090676201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,32,0.0576853354771932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,16384,0.43490134345160586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,8192,0.1629777749379476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,2048,0.2502817842695448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,768,0.0943795575035943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,4096,0.16002933184305826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,2560,0.005548444473081165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,51200,0.40711021423339844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,12288,0.04212799999448988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,768,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,1024,0.04584177666240268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,8192,0.14454933007558188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,1536,0.1307671070098877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,32,0.005145777844720417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,3584,0.006991110742092133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,768,0.004181333300140169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,2560,0.015049777097172208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,1536,0.014670222997665405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,65536,0.42279643482632107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,512,0.00554933316177792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,51200,0.01903733279969957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,7168,0.020972443951500788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,128,0.04119022356139289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,65536,0.16285599602593318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,3584,0.027281777726279363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,2048,0.011400000088744693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,65536,0.12317600515153672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,768,0.0036551112102137674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,1024,0.016171556380059984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,1024,0.034881777233547635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,4096,0.024562666813532513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,4096,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,3072,0.010067555639478896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,2560,1.029927995469835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,10240,0.012096000214417776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,1536,0.010497777826256223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,10240,0.01684711045689053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,768,0.0347999996609158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,16384,0.09295466873380874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,768,0.009979555176364051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,51200,1.0995395448472765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,2048,0.8755608664618598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,3072,0.02753777801990509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,51200,0.13971288998921713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,2048,0.019838222199016146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,10240,0.16287467214796278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,12288,0.08620088630252415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,256,0.019698666201697457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,7168,0.010819555156760745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,2048,0.01036888857682546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,12288,0.07990844382180108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,768,0.005481777919663324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,2048,0.04353422257635328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,1536,0.021606221795082092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,6144,0.05790399842792087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,1536,0.013082666529549493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,768,0.009373333719041612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,512,0.004200889004601372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,128,0.005117333183685939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,512,0.005431111074156231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,51200,0.13480267259809706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,1536,0.011519110865063138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,4096,0.010796444283591377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,3072,0.5565679868062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,1024,0.04817688796255323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,4096,0.02011644509103563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,16384,1.342365370856391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,128,0.013982221484184265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,10240,0.3798897796207004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,1024,0.006421333385838403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,2560,0.020957332518365648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,8192,0.042655110359191895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,6144,0.006572444405820634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,8192,0.4791119893391927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,65536,0.01184622198343277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,1536,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,2048,0.0559448864724901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,4096,0.34430578019883895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,1024,0.00868888861603207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,768,0.00565422202150027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,256,0.0031475555151700974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,51200,0.5345493422614204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,64,0.0034826666944556763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,7168,0.1330151160558065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,3072,0.01036800030204985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,128,0.13474488258361816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,6144,0.4700062010023329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,6144,0.1521031061808268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,6144,0.06167466772927178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,4096,0.01484444406297472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,16384,0.037265777587890625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,4096,0.010617777705192566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,8192,0.015045333239767285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,5120,0.06889155838224623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,65536,0.18668977419535318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,1536,0.00812177773978975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,3072,0.3516311115688748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,65536,0.15480266677008733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,5120,0.014855111638704935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,16384,0.06872177786297269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,512,0.02494755552874671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,256,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,256,0.005614222337802251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,65536,4.29330317179362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,8192,0.01420177850458357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,51200,0.833638244205051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,1536,0.04047733214166429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,3072,0.011579555769761404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,12288,0.07077511151631673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,128,0.0032515554792351196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,64,0.0028719999310043124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,12288,0.06108355522155762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,3072,0.031168000565634832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,2560,0.007324444750944774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,1536,0.00980355590581894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,12288,0.06349511279000176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,4096,0.032066666417651706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,12288,0.04276355438762241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,768,0.02087733315096961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,128,0.013978666729397245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,1024,0.008047999607192146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,1024,0.005924444645643234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,2560,0.011176889141400656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,64,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,3584,0.01074222226937612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,10240,0.021534222695562575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,12288,0.021743999587164983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,1024,0.005869333528810077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,1024,0.006233777850866318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,3072,0.01462311049302419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,3072,0.006970666348934174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,128,0.022848000129063923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,10240,0.01891288823551602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,3584,0.007863110966152614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,128,0.00508355556262864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,2560,0.049504889382256397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,8192,0.045476443237728544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,256,0.006830222076839871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,8192,0.058655997117360435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,256,0.019144000278578866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,256,0.00507733349998792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,65536,2.959809833102756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,12288,0.008211555580298105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,2048,0.0694248874982198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,12288,0.7822746700710721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,64,0.01681777834892273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,128,0.010139555566840703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,512,0.004935111022657818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,2560,0.006108444597986009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,256,0.003564444267087512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,65536,0.9827093548244901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,1536,0.08423999945322673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,128,0.005330666485759947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,512,0.00869777798652649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,8192,0.021239999267790053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,32,0.05184444453981188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,3584,0.16291199790106878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,512,0.006113777971929974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,65536,0.17426844437917074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,8192,0.41252178615993923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,4096,0.010599999792046018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,32,0.006350222147173352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,512,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,6144,0.010080888867378235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,2560,0.021920889616012573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,768,0.005744888964626525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,3072,0.006491555521885554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,128,0.014479999740918478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,8192,0.018264888061417472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,12288,0.027821332216262817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,256,0.024665777881940205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,5120,0.013234666652149625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,6144,0.10766666465335423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,12288,1.1977155473497179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,65536,0.025015999873479206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,2048,0.007904888855086433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,1536,0.005786666853560342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,16384,0.02573866645495097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,10240,0.009441778063774109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,51200,0.019296000401178997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,512,0.008543111383914948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,128,0.0033164444482988785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,256,0.004395555704832077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,6144,1.5180924733479817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,16384,0.04344800114631653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,1024,0.02861777941385905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,8192,0.024475556280877855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,64,0.005499555418888728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,7168,0.12695377402835423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,128,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,3072,1.7916747199164496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,768,0.027980443504121568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,1536,0.016373333003785875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,3584,0.010426667001512317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,256,0.0030942221896515954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,8192,0.023620444867346022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,128,0.009963555468453301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,51200,0.12147200107574463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,16384,0.061448891957600914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,65536,0.07690399885177612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,768,0.021866665946112737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,32,0.1313368876775106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,64,0.019908444748984445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,512,0.00600000015563435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,2048,0.0074746666683091064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,16384,0.008859555754396651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,1024,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,4096,0.10362399948967828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,2048,0.04577155576811897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,51200,1.9316647847493489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,32,0.007487110793590546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,10240,2.4238319396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,128,0.01053955571519004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,51200,1.0187075932820637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,10240,0.4586888949076335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,768,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,2048,0.010518222219414182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,3584,0.01348355578051673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,8192,0.060772445466783315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,4096,1.4475662443372939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,6144,0.018208000395033095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,512,0.006724444528420766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,5120,0.013991110854678683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,51200,0.2762986554039849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,64,0.02346933384736379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,1024,0.007699555820888943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,128,0.010808889236715106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,1024,0.006412444429265128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,1536,0.021103999680942957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,65536,0.12133244673411052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,256,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,2048,0.008262222011884054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,65536,5.4649768405490455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,8192,0.03180088930659824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,4096,0.01253600004646513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,3584,0.010871111518806882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,64,0.006630222416586346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,64,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,256,0.0069520001610120135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,2560,0.006755555669466655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,3584,1.3982648849487305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,16384,0.036454220612843834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,65536,0.050903111696243286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,5120,0.08024622334374322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,256,0.0046880000995265115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,1024,0.008849778109126622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,7168,0.05873777468999227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,1024,0.010749333434634738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,65536,0.014883556299739413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,16384,0.024917332662476435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,10240,1.1621501710679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,3584,0.008967111508051554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,2048,0.014353776971499125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,3072,0.24987644619411895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,256,0.034049779176712036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,256,0.021334222621387903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,1536,0.007231999602582719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,2560,0.008604444563388824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,3072,0.008850666383902231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,64,0.004319111092223061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,1024,0.015172445111804538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,3072,0.03408888975779215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,8192,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,256,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,256,0.00774133370982276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,768,0.003850666599141227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,5120,0.04579377836651272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,12288,0.5434026718139648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,768,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,768,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,16384,0.03642222285270691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,16384,0.06802488697899713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,2560,0.018573333819707234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,5120,0.012000888586044312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,16384,2.00838131374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,128,0.006687110910813014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,32,0.0029760001020299066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,6144,0.27137600051032174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,16384,0.042618665430280894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,64,0.005335111171007156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,768,0.007256888680987888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,16384,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,512,0.0039048890272776284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,4096,0.026946667167875502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,128,0.010932444698280759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,3584,0.06561688582102458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,51200,0.2446613311767578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,51200,0.10807733403311835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,4096,0.15155111418830022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,1536,0.03564888901180691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,1024,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,65536,0.07887644237942167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,7168,0.017779555585649278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,65536,0.1402240064409044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,51200,1.1542008717854817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,2048,0.16325067149268255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,10240,0.039684444665908813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,768,0.009135999613338048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,64,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,1536,0.009308444129096137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,1536,0.017111111018392775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,2560,0.12431822882758246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,12288,0.3734097745683458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,5120,0.15732977125379774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,32,0.007855999800893996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,2560,0.028140445550282795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,1024,0.03609777821434869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,256,0.14006488853030735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,256,0.07796800136566162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,256,0.022738666998015508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,6144,0.04134133458137512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,8192,0.7450275421142578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,16384,0.12224800056881374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,6144,0.02619733413060506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,65536,1.5216621822781031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,1024,0.0583840012550354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,2560,0.008466666771305932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,7168,0.04133866561783685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,32,0.12135556009080674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,1536,0.030228445927302044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,3072,0.023448000351587932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,4096,0.0829004446665446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,10240,0.02370133333735996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,51200,0.15142756038241917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,12288,0.07235288619995117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,768,0.011414222419261932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,5120,0.07337866889105903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,16384,0.09388799799813165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,128,0.01276533305644989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,10240,2.6071554819742837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,3072,0.46040799882676864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,3072,0.01648533344268799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,7168,0.25673866271972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,7168,0.07714222537146674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,32,0.005081777771313985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,32,0.03081866767671373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,512,0.011329777538776398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,2560,0.006383111079533895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,3584,0.067921777566274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,1024,0.021611554755104914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,256,0.03506222367286682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,512,0.021870222356584337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,8192,0.020063999626371596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,51200,1.9645199245876734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,6144,0.012968888713253869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,51200,0.09930933184093899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,256,0.005110222018427319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,768,0.040009776751200356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,3072,0.007553777760929531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,768,0.008303110798199972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,1024,0.1239466667175293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,128,0.0032844444115956626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,2048,0.010439111126793755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,1024,0.021939555803934734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,51200,0.08891733487447102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,7168,0.027708444330427382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,51200,0.008227555288208855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,6144,0.14945867326524523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,256,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,1536,0.029478222131729126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,4096,0.11810399426354302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,4096,0.012305777933862476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,10240,0.03429244293106927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,7168,0.10902933279673259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,65536,1.794742160373264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,5120,0.012834666503800286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,64,0.020163555939992268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,4096,0.021703110800849065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,5120,0.012171555724408893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,2560,0.06243377923965454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,32,0.004917333523432414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,1536,0.010195555786291758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,32,0.021218667427698772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,1536,0.04296888907750448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,768,0.008114666574531132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,1536,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,8192,0.006540444576077991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,7168,0.012682666381200155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,2048,0.01688088807794783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,4096,0.01811822255452474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,2560,0.03449155555831061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,6144,0.3026008870866564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,3584,0.02530577778816223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,5120,0.01865244408448537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,512,0.0659244457880656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,3584,0.024466666910383437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,3584,0.03429066803720262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,2560,0.08626577589246963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,5120,0.21083288722568086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,1024,0.01052977806992001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,3584,0.27043022049797905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,64,0.09344177775912815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,2048,0.035728000932269625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,6144,0.2939377890692817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,5120,0.528521802690294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,16384,0.2550364335378011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,768,0.014661333627170987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,1536,0.024115555816226538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,32,0.07233066691292657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,12288,0.057152887185414634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,12288,0.17238933510250518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,2048,0.042986667818493314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,3072,0.048355556196636625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,512,0.0034222222036785553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,2048,0.023211555348502264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,128,0.0075786668393347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,32,0.003047111133734385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,6144,0.03680977887577481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,12288,0.03608355588383145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,128,0.017867555220921833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,12288,0.04784355560938517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,1536,0.007858666280905405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,512,0.01697688963678148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,8192,0.06408355633417766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,128,0.007079111205206976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,10240,0.02331022255950504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,1024,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,64,0.0027253333893087176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,7168,0.3042008876800537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,51200,4.950275421142578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,256,0.10268088844087388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,4096,0.05518311262130737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,5120,0.025282666087150574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,6144,0.020439111524158053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,16384,0.02979733215437995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,5120,0.016140444411171805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,6144,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,12288,0.06606844398710462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,2048,0.006625777731339137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,1536,0.004767111192146937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,6144,0.4180462095472548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,3584,0.03636888994110955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,8192,0.04413422279887729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,5120,0.016487111647923786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,10240,0.01661422186427646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,64,0.007075555622577667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,512,0.0075146663520071245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,65536,0.08278399705886841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,51200,5.048148261176215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,16384,0.14141421847873265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,1024,0.009943111075295342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,51200,0.016528000434239704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,7168,0.08733510971069336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,10240,0.10047022501627605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,16384,0.1842106713189019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,4096,0.011877333124478659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,2048,0.006265777680608962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,6144,0.027758222487237718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,32,0.016032889485359192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,2560,0.01867822143766615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,128,0.003548444559176763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,4096,0.011569778124491373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,1536,0.02976977825164795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,2048,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,16384,0.027033777700530157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,65536,0.8621493445502387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,8192,0.14139555560217962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,4096,0.02987288766437107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,256,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,3072,0.011713777979214987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,1536,0.041365332073635526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,1536,0.009004444711738164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,5120,0.008634666601816813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,6144,0.02624888883696662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,512,0.020069332586394418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,10240,0.014091556270917257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,65536,0.1093555556403266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,3584,0.010413332945770687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,12288,0.15343556139204237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,3584,0.1291217803955078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,12288,0.19912978013356528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,2560,0.010188444621033138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,1024,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,1024,0.006066666709052191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,768,0.010155555274751451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,2048,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,1024,0.0042133331298828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,32,0.005007110950019625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,65536,1.4313928816053603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,2560,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,2048,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,51200,0.0350560016102261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,2560,0.009888889061080085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,6144,0.7002088758680555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,512,0.015765332513385348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,1536,0.011471110913488599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,64,0.02351288828584883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,5120,0.05681066380606758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,2560,0.02014222244421641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,768,0.008849778109126622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,16384,0.44170223342047793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,5120,0.019167999426523846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,10240,0.27685244878133136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,512,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,16384,0.025757332642873127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,1024,0.00997688869635264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,1536,0.008826666408114964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,6144,0.08633333444595337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,5120,0.020167110694779288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,32,0.005059555586841372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,2048,0.04399111204677158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,1536,0.06520355410046048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,3584,0.16406577163272434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,6144,0.03307377629809909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,7168,0.017861333158281114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,3072,0.029382222228580054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,64,0.031013333135181006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,6144,0.013353777428468069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,1536,0.020591111646758187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,64,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,5120,0.009637333452701569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,51200,0.174563553598192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,16384,0.2963866657680935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,512,0.016044444508022733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,1536,0.012233778006500669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,128,0.0058266665372583605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,768,0.0035128887328836652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,12288,0.8385661972893609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,4096,0.06246311134762234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,10240,0.07971555656856961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,768,0.02513866623242696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,2560,0.014442667365074158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,64,0.006537777682145436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,8192,0.08215022087097168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,16384,0.06693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,10240,0.060692442788018115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,1536,0.007037333316273159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,1024,0.022073777185546026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,65536,1.2294036017523873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,4096,0.021178666088316176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,51200,0.06264622343911065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,768,0.006370666540331311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,768,0.018559111489189994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,1536,0.00610133343272739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,2560,0.029652443197038438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,512,0.02241511146227519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,5120,0.009714666340086196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,5120,0.009730666875839233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,128,0.004095111042261124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,8192,3.383820427788628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,5120,0.25126579072740346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,256,0.030680888228946265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,128,0.009730666875839233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,16384,0.20711644490559897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,32,0.04618488748868307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,3584,0.09871021906534831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,4096,0.046894222497940063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,1024,0.008672888908121321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,1536,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,2560,0.49530310100979275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,128,0.00574933323595259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,1536,0.006537777682145436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,7168,0.012230222423871359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,256,0.011918221910794577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,64,0.019502222537994385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,256,0.009692444569534725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,32,0.08862577544318305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,1024,0.0090515555606948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,1536,0.008807111117574904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,3072,0.01994044416480594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,128,0.005066666752099991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,16384,0.2961680094401042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,256,0.016736888223224215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,3584,0.017294221454196505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,3072,2.5977350870768228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,4096,0.023709333605236475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,5120,0.02438844409253862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,1536,0.010685332947307162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,3072,0.03233244352870517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,6144,0.16091644763946533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,3072,0.013889777991506787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,5120,0.011634666886594562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,4096,0.020611556039916146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,12288,0.03090133269627889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,12288,0.03907910982767741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,2560,0.011772444678677453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,65536,0.12415733602311875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,512,0.01714666684468587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,64,0.0027840000887711844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,3072,0.6143760151333278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,1024,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,2560,0.15160355303022596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,128,0.17293688986036512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,3584,0.1862106720606486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,2560,0.007481777833567725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,65536,0.20950755808088514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,64,0.005252444495757421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,51200,2.3903412289089627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,1024,0.06636889113320245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,768,0.006566222343179915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,8192,14.091830783420138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,32,0.010665777656767102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,128,0.006728000111050076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,51200,0.12347466415829128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,32,0.008010666403505538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,65536,0.08542399936252171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,10240,0.4877617624070909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,64,0.005944889038801193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,3072,0.010416888528399997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,16384,0.11158043808407253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,1024,0.0058168888919883305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,128,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,1536,0.11053422424528335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,3072,0.008579555484983657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,128,0.05795288748211331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,3072,0.023761777414215937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,3072,0.01628533336851332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,3072,0.022364444202846948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,12288,0.05589333507749769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,64,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,2560,0.01682044400109185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,4096,0.014583110809326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,128,0.003954666770166821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,64,0.0026542221506436667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,8192,0.12602489524417454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,65536,0.01275111072593265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,512,0.021243555678261652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,7168,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,5120,0.02390755547417535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,4096,0.009543110926946005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,512,0.00453244439429707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,1024,0.10867644680870904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,3584,0.012137778103351593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,6144,0.015477332803938123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,4096,0.026968000663651362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,6144,0.01590222285853492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,4096,0.01593155496650272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,256,0.00301155551440186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,12288,0.05186133252249824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,768,0.3046311007605659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,51200,0.019039111004935372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,32,0.007733333441946242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,6144,0.012220444778601328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,8192,0.03404355711407132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,256,0.03500355614556207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,1536,0.007460444337791867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,16384,0.8677751223246256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,2048,0.011601777540312873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,128,0.006973333656787872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,64,0.0032799999333090256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,3584,0.04042311178313361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,2560,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,768,0.005715555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,768,0.7715902328491211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,8192,0.059439109431372754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,1024,0.12439733081393772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,16384,0.7766568925645617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,16384,0.19805688328213164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,3072,0.019095111224386428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,256,0.005631999837027655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,3584,0.43813154432508683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,5120,0.05330666568544176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,4096,0.0063742221229606206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,16384,0.019048889478047688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,2048,0.012133333418104382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,2048,0.28730400403340656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,7168,0.13002578417460123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,128,0.007310222420427535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,16384,0.10484000047047932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,128,0.004807999978462855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,32,0.005023111071851519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,65536,3.0713689592149525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,6144,0.03844622108671401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,65536,0.9217928780449761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,768,0.005193777796294955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,1024,0.005112888912359874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,512,0.015088889333936902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,51200,0.0631200008922153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,7168,0.195360885726081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,3584,0.009944000177913243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,768,0.019818666908476088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,128,0.006098666538794835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,1024,0.035990221632851496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,3072,0.039423998859193586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,512,0.02717511190308465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,3072,0.008231110870838165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,65536,0.012480888929631976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,7168,0.007221333682537079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,4096,0.3686337735917833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,12288,0.09477688868840535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,512,0.035282668140199445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,256,0.007826666865083907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,64,0.062202665540907115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,51200,2.839768091837565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,768,0.016016888949606154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,8192,0.17989244725969103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,65536,0.4113973246680365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,51200,0.6662435531616211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,3584,0.013647110925780402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,512,0.015454222758611044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,3584,0.008924444516499838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,512,0.030544890297783747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,4096,0.028832889265484277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,65536,0.15309955014122859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,256,0.012576888832781049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,2048,0.006511999915043513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,10240,0.017350221673647564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,128,0.008636444807052612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,2048,0.013444444371594323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,768,0.008034666379292807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,10240,0.22474400202433267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,2560,0.016364443633291457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,3584,0.04833333359824287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,768,0.005693333430422678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,7168,0.013688888814714221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,65536,1.4412719938490126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,10240,0.03769244419203864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,7168,0.02959733208020528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,3072,0.10406933228174846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,4096,0.010694222317801582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,3584,0.010155555274751451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,12288,0.083024885919359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,6144,0.035338666703965925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,16384,0.024863999750879075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,6144,0.036079999473359846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,51200,2.7140543195936413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,51200,0.1321857770284017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,7168,0.01367022262679206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,8192,0.011766222616036734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,512,0.00370133337047365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,65536,0.10118310981326634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,256,0.2325493229760064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,4096,0.09363289011849298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,2560,0.06439644760555692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,2048,0.045351111226611666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,512,0.010293333066834344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,2048,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,32,0.0029155556112527847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,12288,0.8993458218044705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,1024,0.01625777781009674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,256,0.007816000117195977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,256,0.024518221616744995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,12288,0.31348710589938694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,8192,0.020796444680955675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,5120,0.14161600006951228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,6144,0.0333262218369378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,512,0.026495110657480028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,10240,0.2504711151123047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,8192,0.012153777811262341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,768,0.020672889219390023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,5120,0.009900444083743626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,64,0.009022222624884712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,32,0.027659555276234944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,768,0.007999111380841997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,12288,28.838619656032986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,3584,0.007536888950400882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,10240,0.012319111161761813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,768,0.11888444423675537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,2048,0.006907555378145642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,64,0.005139555368158552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,10240,0.11116622553931342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,64,0.005584888988071018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,1536,0.0744311081038581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,10240,0.03914489017592536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,12288,0.05408089028464424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,1024,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,8192,0.022685334086418152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,1024,0.003911111089918348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,4096,0.017465778523021273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,4096,0.07560800181494819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,7168,0.03016888764169481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,1536,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,32,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,2048,0.01516266663869222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,65536,1.6670026779174805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,5120,0.10848889085981582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,128,0.008119111259778341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,10240,0.01826400061448415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,2048,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,51200,0.05866310993830363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,5120,0.07272622320387098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,3584,0.2421813276078966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,51200,0.09934844573338826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,5120,0.013052444491121503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,3072,0.03662755423121982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,12288,0.4030284351772732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,768,0.004303110970391167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,65536,3.5040880839029946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,8192,0.048304888937208385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,64,0.0029324444217814338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,512,0.007974222302436829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,2048,0.007756444315115611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,256,0.0033644443998734155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,64,0.0029475556479560006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,10240,0.9553030861748589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,65536,6.207529703776042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,65536,0.7648346689012316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,2048,0.008187555604510838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,3072,0.007277333074145847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,1024,0.24071110619439018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,12288,0.07260888814926147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,5120,0.016806221670574613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,6144,0.35752354727851016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,1536,0.004680888934267892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,32,0.005415110952324337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,3584,0.01628977722591824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,128,0.0033173333439562055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,3072,0.023835556374655828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,6144,0.02181600034236908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,10240,1.4924400117662218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,12288,0.02275466587808397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,7168,0.01534577707449595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,6144,0.007530666887760162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,1024,0.013281777501106262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,4096,0.024002666274706524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,7168,0.01610133383009169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,16384,0.15721600585513645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,3072,0.3190026548173693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,6144,0.03542933199140761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,8192,0.019896888070636325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,10240,0.017101332545280457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,1024,0.009275555610656738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,2048,0.06911289029651217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,3072,0.015625778171751235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,5120,0.018236445056067575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,64,0.007287110719415877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,10240,3.6803432040744357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,51200,1.0685475667317708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,512,0.023016000787417095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,64,0.003909333298603694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,7168,36.40876261393229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,2048,0.11859556039174397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,10240,0.5408399899800619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,2560,0.01608444419172075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,16384,1.468631108601888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,32,0.003510222252872255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,4096,0.04413777920934889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,1536,0.012582222620646158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,1536,0.02716711163520813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,12288,0.04686489038997226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,1536,0.2833102279239231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,7168,0.06772622134950426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,65536,7.66158209906684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,1536,0.007108444141017065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,2048,0.008827555510732863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,51200,2.5599208407931857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,1536,0.006024888820118374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,1024,0.00942755573325687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,256,0.0069715554515520734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,3072,0.48659467697143555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,2048,0.05787999762429131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,65536,1.0518417358398438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,2048,0.14960532718234593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,128,0.025612443685531616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,8192,0.03712000118361579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,5120,0.014965333872371249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,256,0.010637332995732626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,4096,0.2035813331604004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,32,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,256,0.006395555618736479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,4096,0.011572444604502784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,2560,0.014291556345091926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,128,0.011360000405046674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,65536,2.20704714457194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,16384,0.04350044329961141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,51200,0.3284577793545193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,1024,0.010798222488827176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,12288,0.021984888447655573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,6144,0.010231999887360467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,7168,0.09950221909417047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,1536,0.031332443157831825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,32,0.005392000079154968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,12288,0.013368888861603208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,256,0.00591111100382275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,16384,0.030593779351976182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,256,0.002999111182159848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,32,0.008799999952316284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,4096,0.43225778473748105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,256,0.006212444355090459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,3584,0.07040266858206855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,2048,0.05610489183002048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,3584,0.037083556254704796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,32,0.0027866665687825945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,12288,0.024448888169394598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,1024,0.009870222045315636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,2048,0.16395200623406306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,64,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,128,0.004906666775544484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,4096,0.05123733480771383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,5120,0.562181313832601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,5120,0.15459733539157444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,8192,0.036305778556399874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,65536,6.64701165093316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,4096,3.3449537489149304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,7168,0.017281777328915067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,51200,0.1447004477183024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,3072,0.6969422234429253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,7168,0.014736000034544202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,12288,0.09062221977445815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,2560,0.010499555203649731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,7168,0.016826666063732572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,6144,0.014367999302016364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,1024,0.05270488725768196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,10240,0.25365956624348956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,8192,3.197289784749349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,3584,0.21214400397406685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,3072,0.0643706652853224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,512,0.024453333682484094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,65536,0.26251289579603404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,3072,0.05254400107595655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,5120,0.2658257749345568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,8192,0.19003110461764863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,51200,0.0640853312280443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,3072,0.014397333065668741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,65536,0.08017333348592122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,128,0.004860444615284602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,5120,0.01760088900725047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,2048,0.007126222054163615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,256,0.005095110999213325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,4096,1.1099164750840929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,6144,0.2563680013020833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,256,0.011192888849311404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,768,0.18964000542958578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,3072,0.00796088907453749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,512,0.03233688738611009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,3584,0.014803555276658801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,65536,0.19722399446699354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,7168,0.015406222807036506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,512,0.004395555704832077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,256,0.005473777651786804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,768,0.02570666703912947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,8192,0.022543110781245764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,2560,0.01089333328935835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,51200,0.19197244114345977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,5120,0.10723911391364203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,8192,0.11786577436659072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,4096,0.12431466579437256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,16384,0.33872177865770126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,51200,0.28761066330803764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,12288,0.17464000648922393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,768,0.0331928895579444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,6144,0.010099555883142684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,10240,0.039317333035998873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,8192,0.20851643880208334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,4096,0.02128355536195967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,7168,0.012135110795497894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,2560,0.043261332644356616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,128,0.0177831103404363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,10240,0.03253066539764404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,256,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,7168,0.0188044442070855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,64,0.014967110421922473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,10240,0.3007982307010227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,8192,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,7168,0.6174542109171549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,768,0.012630222572220696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,256,0.0046693334976832075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,256,0.0038399998512532976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,1536,0.009956444303194681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,32,0.00814222213294771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,256,0.0514302220609453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,16384,0.16550400522020128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,51200,0.07269066572189331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,2560,0.26818932427300346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,7168,0.10800000031789143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,512,0.011905777785513135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,16384,1.0547893312242296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,256,0.006047111004590988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,256,0.012224888636006249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,6144,0.07208000289069282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,10240,0.21789244810740152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,1536,0.0069146665434042616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,3584,0.01239822225438224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,2560,0.011369778050316704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,2048,0.014538667268223233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,6144,0.03795999950832791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,16384,0.04855733447604709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,1536,0.06005422274271647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,512,0.016858667135238647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,1536,0.027729777826203242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,2048,0.018736888964970905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,3072,0.01218844453493754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,2560,0.007115555306275685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,32,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,12288,0.04501333170466953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,512,0.04481955700450473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,10240,0.03611733184920417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,3584,0.047832889689339526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,3072,0.008227555288208855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,12288,0.2226817872789171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,2048,0.049808889627456665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,2048,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,8192,0.026887999640570745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,7168,0.03203466534614563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,16384,0.05493333273463779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,3072,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,512,0.0033902221669753394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,5120,0.05907999806933933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,1024,0.005107555538415909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,51200,0.054367111788855664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,7168,0.10686044560538398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,6144,0.014576888746685453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,2048,0.021407999926143225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,6144,0.02920888861020406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,12288,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,3584,0.07019466824001737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,64,0.005677333308590784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,8192,0.12488799624972874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,65536,4.592286427815755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,3072,0.06807821989059448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,65536,0.414631101820204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,65536,5.002505832248263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,64,0.0030862221287356485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,10240,0.012396444049146442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,128,0.005139555368158552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,256,0.006709333509206772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,10240,0.16383643945058188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,1536,0.006912888752089606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,7168,1.156582196553548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,2560,0.07855200105243258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,768,0.00889155517021815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,2560,0.0058959999846087555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,51200,3.6530905829535594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,8192,0.2701084348890516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,51200,1.3337217966715496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,2048,0.08184088601006402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,3072,0.21577600638071695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,12288,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,16384,0.042209777567121715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,51200,5.195214165581597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,4096,0.02048000031047397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,7168,0.021762667430771723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,12288,0.014214222629865011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,6144,0.04641688863436381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,1536,0.018072888255119324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,4096,0.01428711083200243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,7168,0.02052977846728431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,1024,0.10708977778752644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,6144,0.038373334540261164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,4096,0.03501155641343858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,5120,0.1883066627714369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,3072,0.08355110883712769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,512,0.016160888804329764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,12288,0.053780443138546415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,2048,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,3584,0.006870222174459034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,5120,0.17598844899071586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,128,0.011685333318180509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,4096,0.019111111760139465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,1536,0.0063279999627007385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,51200,1.8797253502739801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,6144,0.015229332778188916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,2560,2.069183137681749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,7168,0.02032355633046892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,4096,0.019376888871192932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,12288,2.9389635721842446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,4096,0.02149155570401086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,12288,0.018752889500723947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,3584,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,1536,0.13375643889109293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,51200,1.0027155346340604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,1536,0.00527022240890397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,5120,0.01679111189312405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,7168,0.31241599718729657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,4096,0.019135110908084445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,2560,0.08084266715579562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,128,0.005058666484223472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,7168,0.2553155687120226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,7168,0.06193066967858208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,1024,0.03591555688116286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,1024,0.010871999793582492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,1024,0.041088889042536415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,32,0.005008888741334279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,1536,0.01661155621210734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,6144,0.11610133118099636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,1536,0.012540444731712341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,512,0.005484444399674733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,768,0.025802666942278545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,16384,0.04900622367858887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,16384,31.305633544921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,256,0.023102222217453852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,768,0.028359111812379625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,6144,0.021664000219768945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,51200,1.4797013600667317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,512,0.24125154813130698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,16384,3.7780685424804688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,5120,0.26367465655008954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,8192,0.34714044464959043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,2560,0.007401777638329401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,12288,0.31571467717488605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,3584,0.1970693270365397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,64,0.00499288903342353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,3584,0.09432978100246853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,6144,0.011995555626021491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,6144,0.1615875562032064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,7168,0.13352978229522705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,6144,0.027262222435739305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,128,0.004454222404294544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,10240,0.018055111169815063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,6144,0.0064586665895250105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,1024,0.008105778031879002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,5120,0.019550222489568923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,5120,0.007616889145639207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,256,0.004383111165629493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,512,0.12436266740163167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,1536,0.006745777610275481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,7168,0.013192888763215808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,1536,0.439355558819241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,4096,0.03501866592301263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,2560,0.22249955601162383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,5120,0.18330132961273193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,12288,0.023341332872708637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,2560,0.10622400045394897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,10240,0.02698755595419142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,768,0.06603378057479858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,3072,0.015290665957662793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,512,0.036630223194758095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,768,0.007870222131411234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,5120,8.485666910807291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,7168,0.059980445437961154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,10240,0.02827377782927619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,128,0.009092444347010719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,8192,0.17894311745961508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,2560,0.09230489201015896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,5120,0.06819999880260892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,768,0.0036319999231232535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,2560,0.01331733332739936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,1024,0.006069333189063602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,8192,0.023396443989541795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,1536,0.013456888496875763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,32,0.008481777376598781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,65536,0.2836746639675564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,5120,0.012449777788586087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,1536,0.004563555535342959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,51200,2.015742195977105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,32,0.003342222215400802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,51200,0.23040265507168242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,768,0.005718222094906702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,6144,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,10240,0.3987751007080078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,7168,0.007567110988828871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,2560,0.014563555518786112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,128,0.004848888764778773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,768,0.020599999361568026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,512,0.03686666819784377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,12288,0.04205866654713949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,12288,0.03991377684805129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,1536,0.0691208905643887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,6144,0.01385600037044949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,7168,3.5974693298339844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,512,0.06609777609507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,6144,0.032552000549104475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,6144,0.060512887107001416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,3584,0.006913777854707506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,64,0.01686844395266639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,256,0.005695999910434087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,3584,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,6144,0.870508458879259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,128,0.007034666836261749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,2048,0.024672889047198828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,2048,0.017302221722073026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,8192,0.038931555218166776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,3072,0.012919999659061432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,512,0.008074666890833113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,2048,0.07225688960817125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,51200,0.11065422164069282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,5120,0.016639111770523917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,5120,0.6880355411105685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,1536,0.03200622068511115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,2560,0.006625777731339137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,3584,0.012216889195972018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,65536,3.2403288947211375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,51200,1.8418142530653212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,6144,1.6055244869656031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,1536,0.00814488861295912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,10240,0.029640889830059473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,1536,0.008714666797055138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,7168,0.02047288914521535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,768,0.004853333450025982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,3584,0.1723804473876953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,51200,0.05603555838267008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,10240,0.04916444420814514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,3584,0.021431111627154883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,1536,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,2560,0.02953333324856228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,128,0.012727999852763282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,64,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,4096,0.012629333469602795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,1024,0.011744000017642975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,32,0.007035555938879649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,4096,0.017180444465743173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,12288,0.13767466280195448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,51200,0.1049484411875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,1536,0.06259555286831327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,2560,0.021778666310840186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,128,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,3584,0.00756177802880605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,12288,0.2374995549519857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,2048,0.005320000151793162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,6144,0.1200737820731269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,6144,0.09191377957661946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,65536,0.2129297786288791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,2048,0.16575200027889675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,5120,0.020118221640586853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,12288,0.010419555836253695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,64,0.0032497776879204642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,2048,0.013755555782053204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,256,0.014763555592960782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,8192,0.06457244687610202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,1536,0.014940443966123792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,64,0.05640000104904175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,256,0.021689777572949726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,512,0.005897777775923411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,16384,0.056120885743035205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,4096,0.029855999681684706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,128,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,2560,0.00925244473748737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,6144,0.09588177998860677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,1536,0.02973244587580363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,4096,0.011125333607196808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,768,0.7601368692186145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,4096,0.13557333416408962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,64,0.006829333388143116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,1536,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,256,0.007271111011505127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,10240,0.14621510770585802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,2560,0.007248000138335758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,8192,0.1140364408493042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,1024,0.0039822223285834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,256,0.004384888956944148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,2560,1.4208080503675673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,3584,0.034605334202448525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,3072,0.018144889010323417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,4096,0.021026665965716045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,1024,0.007054222126801808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,65536,2.2465271419949002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,8192,0.03731288843684726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,8192,0.026151110728581745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,3072,0.044610665904151074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,5120,0.014111111561457316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,2560,0.017696888910399545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,1024,0.014594667487674289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,16384,0.09457866350809734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,3584,2.520477294921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,51200,1.448111957973904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,128,0.004569777597983678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,3072,0.008777778181764815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,10240,0.09440800216462876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,128,0.0026666666898462507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,32,0.005052444421582752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,256,0.04565777712398105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,1024,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,16384,0.11976977189381917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,16384,0.3117991023593479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,51200,1.6435404883490667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,10240,1.6086284849378798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,3072,0.03453866640726725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,32,0.005111111121045219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,65536,0.09329777956008911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,4096,0.11422488424513076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,64,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,2048,0.005585777676767773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,768,0.034372445609834455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,1536,0.4227946599324544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,16384,0.01496088835928175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,12288,0.12489688396453857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,10240,0.01586666703224182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,768,0.016490666402710807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,8192,0.11752355098724365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,2048,0.06578844123416476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,3584,0.011053333679835001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,512,0.029473778274324205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,10240,0.015556444724400839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,7168,2.589773390028212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,2048,0.02409066590997908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,6144,0.11511466238233779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,5120,0.006582222051090664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,256,0.0033902221669753394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,1024,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,5120,0.22778399785359701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,4096,0.2697804503970676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,3584,0.2848435507880317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,7168,0.015961777832772996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,7168,0.6512622303432888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,12288,0.29268089930216473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,3584,0.02453511125511593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,768,0.015861334072219003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,32,0.02071199980047014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,768,0.0042204442951414315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,6144,0.013798221945762634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,2560,0.015036443869272867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,16384,0.052763554784986705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,12288,0.3557697667015924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,3072,0.022557333111763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,3584,0.020616888999938965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,5120,0.010957332948843638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,1536,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,64,0.010813333094120026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,5120,0.011792889071835412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,65536,0.3381182352701823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,2560,0.007431999676757389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,10240,0.6655679808722602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,3584,0.013367999758985309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,5120,0.2236106660630968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,1536,0.013965333501497904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,51200,0.07036533620622423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,10240,1.0029840469360352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,4096,0.01036888857682546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,5120,0.046458668178982206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,64,0.0034640000926123727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,2048,0.009764444496896531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,512,0.005508444375462002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,64,0.029310223129060533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,65536,0.13539466593000624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,256,0.005199999858935674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,128,0.009322666459613377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,512,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,51200,1.4462791019015844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,65536,5.149787478976779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,10240,0.04225066635343763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,1536,0.01037688884470198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,16384,0.00924444446961085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,3072,1.1277404361300998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,65536,0.9573538038465711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,768,0.03973333372010125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,16384,0.39200178782145184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,4096,0.4067262278662787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,51200,0.030159112479951646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,32,0.06122577852673001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,5120,0.01457244488928053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,768,0.02399377855989668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,32,0.0026977778308921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,6144,0.010063110954231685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,2048,0.07340266969468859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,4096,0.010699555277824402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,2560,0.023079110516442194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,512,0.22733778423733184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,12288,0.1577786604563395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,768,0.01591022147072686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,10240,0.14582488271925184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,128,0.07393422391679552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,128,0.1412995523876614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,8192,0.017519111434618633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,1536,0.0064399999876817065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,3072,0.046738667620552905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,4096,0.035572446054882474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,1536,0.11317066351572673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,32,0.010017777482668558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,128,0.007633777956167857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,1536,0.009841777384281158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,5120,0.009454222189055549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,5120,0.07574844360351562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,10240,0.07366133398479886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,10240,0.4134399890899658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,1024,0.014549333188268872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,5120,0.0993173321088155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,65536,0.024225778049892847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,3584,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,10240,0.008662222160233391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,3072,0.016536000702116225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,32,0.002491555487116178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,51200,0.9178888532850477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,3072,1.9727431403266058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,6144,0.10320978032218085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,65536,0.010203555226325989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,2560,0.016013332539134555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,256,0.01144088887506061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,32,0.008071111308203803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,7168,0.012261333564917246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,5120,0.0074337778819931885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,5120,0.01912088857756721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,8192,0.021662222014533147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,32,0.004262222184075249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,51200,11.002839830186632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,5120,0.25128889083862305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,10240,0.14815733167860243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,768,0.006285333385070165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,1536,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,32,0.003149333306484752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,256,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,3584,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,2048,0.010682666467295753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,51200,0.41714223225911456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,4096,0.008284444610277811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,5120,0.22803110546535918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,8192,0.2780604362487793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,3584,0.014854222536087036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,4096,0.3137297895219591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,10240,2.4705333709716797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,2560,0.017052445146772597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,768,0.018301332990328472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,12288,1.1117209328545463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,2048,0.09462400277455647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,512,0.02490577763981289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,2048,0.01633155511485206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,12288,0.48455201254950625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,16384,0.007143110864692264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,8192,0.062408890989091664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,7168,7.353166368272569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,6144,0.4585439893934462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,8192,0.020614221692085266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,10240,0.023334221707450017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,6144,0.01514933341079288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,1024,0.005686222265164058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,512,0.0886933339966668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,768,0.0258568889564938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,64,0.2311884562174479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,2560,0.008585777547624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,2560,0.007791111038790808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,65536,0.3246684339311388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,768,0.01479111115137736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,7168,0.038945777548684016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,1024,0.008793777889675563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,1536,0.013020444247457715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,16384,1.4361741807725694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,16384,1.304324467976888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,2560,0.08705777592129177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,64,0.008678221868144141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,7168,0.3901048766242133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,51200,0.10949244764116074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,12288,0.03186577889654372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,51200,0.7392177581787109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,64,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,4096,0.3985439936319987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,5120,0.020024889045291476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,6144,0.15166221724616155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,1024,0.011306666665607028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,1024,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,10240,0.23714311917622885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,10240,0.04745066828197903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,2048,0.15417067209879556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,3072,0.01576622161600325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,8192,0.12623377641042074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,65536,0.11411910586886936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,51200,5.7018309699164496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,1536,0.009852444132169088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,65536,0.5503066910637749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,6144,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,512,0.00566133318675889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,4096,0.02139733400609758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,2560,0.04669777883423699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,12288,0.01979022224744161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,6144,0.024710221423043147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,12288,0.3841999901665582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,1024,0.17576177914937338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,4096,0.07021155622270372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,2560,0.03743822375933329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,4096,0.023771555887328252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,6144,0.01556622154182858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,512,0.005585777676767773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,1024,0.020205333828926086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,5120,1.019939528571235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,10240,0.04577689038382637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,6144,0.05173600051138136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,65536,0.3051724433898926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,2048,0.27654488881429035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,10240,0.07058488660388522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,16384,0.026308443811204698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,32,0.06977688603931002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,6144,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,51200,0.0969582200050354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,5120,0.07815821965535481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,8192,0.011653333074516721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,1536,0.009773333039548662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,65536,0.6817262437608508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,7168,0.011152889165613385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,2048,0.1840346654256185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,3072,0.007158222297827403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,4096,0.08112355735566881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,12288,0.1191706657409668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,65536,6.1797972785101996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,1024,0.005489777773618698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,256,0.0047208890318870544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,768,0.0074346669846110875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,12288,0.017991110682487488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,3072,0.0074444446298811175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,65536,0.28490932782491046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,16384,0.2211022244559394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,256,0.03427733315361871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,512,0.004354666504595014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,256,0.006680000159475539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,12288,0.030633777379989624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,8192,0.19082310464647081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,256,0.026917333404223125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,5120,0.009943111075295342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,1024,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,3072,0.008182221816645728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,128,0.0028000000036425064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,65536,0.17181600464714897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,128,0.018185777796639335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,10240,0.11704888608720566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,8192,0.018749333090252347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,12288,0.045175111956066556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,4096,0.13336533970303005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,8192,0.3271991146935357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,51200,0.5200746854146322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,65536,0.06971822182337443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,256,0.0031591111587153543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,3584,0.043544888496398926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,64,0.002445333326856295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,2560,0.019872888922691345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,10240,0.019463999403847587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,8192,0.01405244403415256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,1536,0.027323555615213182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,2560,0.007330666813585494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,3584,0.18176266882154676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,768,0.0035955554081334006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,16384,0.13221866554684109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,256,0.009196444518036312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,51200,0.2651475535498725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,2560,0.017675555414623685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,2560,0.008030222521887885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,6144,0.011804444094498953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,3072,0.0074515557951397365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,65536,4.546257866753472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,1536,0.007834666305118138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,4096,0.009493333597977957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,64,0.003491555651028951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,3072,0.01368533323208491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,128,0.04050755500793457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,2560,0.015301333533393012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,1536,0.006301333506902059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,51200,4.834779527452257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,65536,0.24296622806125215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,1024,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,51200,1.2853982713487413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,7168,0.1698862181769477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,512,0.006688000013430913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,1024,0.004383111165629493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,12288,0.15642933050791422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,7168,0.3881866667005751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,10240,0.08073244492212932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,3072,0.0069253332912921906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,256,0.027958220905727808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,1024,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,10240,0.016136000553766888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,256,0.003456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,2560,0.09440444575415717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,3584,0.010223111344708337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,2560,0.02722666660944621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,1024,0.028416888581381902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,12288,0.05457511213090685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,2560,0.009591110878520543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,10240,0.013547555440002017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,65536,0.3529386785295274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,7168,0.23739555146959093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,5120,0.046893331739637584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,65536,1.7360959582858617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,10240,0.016183111402723525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,1536,0.03266577919324239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,12288,0.020008888509538438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,16384,0.027946667538748846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,10240,0.006798222247097228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,3072,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,12288,0.018217777212460835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,64,0.0029395555870400537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,4096,0.010651555326249866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,3584,0.10796266794204712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,51200,0.22980621125963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,65536,0.07252799802356295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,12288,0.04151022103097703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,10240,0.026143110460705225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,1024,0.009311999711725447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,10240,0.013568888935777875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,51200,0.00998755544424057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,1536,0.029955555995305378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,8192,0.014868444866604276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,512,0.01847555571132236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,4096,0.17760088708665636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,1536,0.03966844413015578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,512,0.003549333247873518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,512,0.004029333177540037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,51200,0.4897919760810004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,12288,0.09674844476911758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,12288,0.027520888381534155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,51200,0.06835555367999606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,7168,0.04215733210245768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,128,0.0028159999185138275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,3584,0.014583999911944071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,16384,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,1536,0.008797333472304875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,5120,0.1494942241244846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,128,0.2591662142011854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,10240,0.01869777838389079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,32,0.003012444410059187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,10240,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,768,0.007347555624114142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,512,0.0034542222403817703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,7168,0.03470666540993585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,8192,0.09789422485563491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,256,0.3265084425608317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,3072,0.010677333507272931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,256,0.003378666730390655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,1024,0.008037333687146505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,768,0.04946133494377136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,64,0.04296088880962796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,8192,0.07353244225184123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,256,0.005144888742102517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,3072,0.012540444731712341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,1024,0.006361777583758037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,3072,0.00998933364947637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,10240,0.10465777582592434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,10240,1.2171573638916016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,51200,0.16517333189646402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,1024,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,1536,0.04903822143872579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,16384,0.027416000763575237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,6144,0.20260622766282824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,51200,0.06862666871812609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,32,0.00739822205570009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,2560,0.02473066747188568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,1024,0.04503555430306328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,32,0.12534844875335693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,32,0.007136888802051544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,6144,0.9274639553493924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,4096,0.01917866700225406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,7168,0.020190222395790946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,128,0.0032248888164758682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,8192,0.029937777254316542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,8192,5.845367007785374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,65536,0.09182755814658271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,12288,0.020703110429975722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,3072,0.022304889228608873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,1024,0.017500445246696472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,5120,0.008903111020723978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,65536,0.3726515505048964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,16384,0.07916266388363309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,16384,0.11482044061024983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,768,0.030716445710923936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,6144,0.024002666274706524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,32,0.0783546699417962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,1536,0.31601066059536403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,12288,0.08913689189487034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,51200,0.5699102083841959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,2560,0.00909511082702213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,1024,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,3072,0.018961777289708454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,2048,0.009827555881606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,8192,0.07190489106708102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,3584,1.1602328618367512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,7168,0.738193776872423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,51200,0.11132710509830052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,128,0.0033848887930313745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,128,0.008006222546100616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,3584,0.03977866636382209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,10240,0.24679909812079537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,768,0.019620445039537217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,12288,0.05001599921120537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,51200,0.28265510665045845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,512,0.028964443339241877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,5120,1.2519991132948134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,64,0.003936888857020272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,3072,0.01817244456874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,2560,0.07199022505018446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,768,0.07028533352745904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,3584,0.13415911462571886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,6144,0.006559999866618051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,256,0.003071111109521654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,4096,2.4832515716552734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,2048,0.01904266741540697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,51200,2.090209749009874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,10240,0.12429332733154297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,256,0.003265777809752358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,1024,0.006812444577614467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,1024,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,4096,0.04503733250829908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,2560,0.010135999984211391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,1536,0.039792001247406006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,256,0.0046817776229646476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,1024,0.07214666737450494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,8192,0.06189511219660441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,16384,0.2847626739078098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,1024,0.027013333307372198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,6144,0.01943911115328471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,128,0.03506044546763102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,32,0.011559111376603445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,10240,0.0588959985309177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,1536,0.012216000093354119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,10240,0.37256267335679794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,64,0.027824888626734417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,64,0.0040693332751592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,65536,0.11044889026217991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,16384,0.00865955568022198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,2560,0.07688355445861816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,3072,0.4703369140625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,5120,0.011788444386588203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,16384,0.05946755409240723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,768,0.015643555257055495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,1536,0.11754843923780654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,512,0.018339556124475267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,1536,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,7168,0.027574222948816087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,5120,0.1576142178641425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,65536,0.36903733677334255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,6144,0.05767911010318332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,128,0.0028133332315418455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,64,0.06543733014000787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,12288,0.01688888834582435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,10240,0.06700533628463745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,4096,0.06248444318771362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,8192,0.8350106875101725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,51200,9.972501966688368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,128,0.006241777704821692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,512,0.005664888769388199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,8192,0.013605333036846586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,1024,0.024737776981459722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,2048,1.6570213105943468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,8192,0.14608089129130045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,10240,0.08357244729995728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,3072,0.012452444268597497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,5120,0.012652444342772165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,1536,0.4286702209048801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,3072,0.48565954632229275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,8192,0.03456177645259433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,8192,0.05876266956329346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,51200,0.08453689018885295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,65536,20.981360541449654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,16384,0.0737324423260159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,8192,0.00868000007337994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,7168,0.12656711207495794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,1536,0.00701422244310379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,12288,0.03741066654523214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,51200,2.2107821570502386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,3072,0.6713840166727701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,3072,0.3280746671888563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,8192,0.12665332688225642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,2560,0.015592000550693937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,2048,0.08170044422149658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,256,0.013184888495339287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,128,0.002805333377586471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,4096,0.042261332273483276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,1536,0.2595884535047743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,256,0.005140444470776452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,768,0.018220444520314533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,10240,0.051106668180889554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,3072,0.012811555630630918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,8192,0.34664000405205625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,3584,0.010832000109884473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,7168,0.05317066775427925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,512,0.015948444604873657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,6144,0.13032710552215576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,2048,0.01108000013563368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,6144,0.05490488807360331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,1536,0.007116444408893585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,2560,0.01771111124091678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,2048,0.01588888963063558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,16384,0.8124471240573459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,5120,0.10135644674301147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,51200,0.9093511369493273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,2048,0.06990133391486274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,1024,0.027457776996824477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,51200,0.1162453360027737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,5120,0.013590222431553734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,128,0.16255111164516872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,256,0.01519733336236742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,3072,0.009741333623727163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,7168,0.009272000028027428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,32,0.013884444203641681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,16384,0.039220445685916476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,4096,0.12156355381011963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,256,0.005033777819739448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,5120,0.009287111461162567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,12288,0.43898489740159774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,5120,0.012831110921170978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,2048,0.007856888903511895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,8192,0.016759999924235873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,1536,0.0110871113008923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,3584,0.08939733107884724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,51200,0.0603804455863105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,3072,0.018368000785509746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,256,0.05494933327039083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,1536,0.005524444497293896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,51200,0.18278222613864473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,3584,0.016209777858522203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,512,0.0046693334976832075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,7168,0.021003555920388963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,51200,0.055564443270365395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,65536,0.3354497750600179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,256,0.0032675556010670135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,7168,0.3326328860388862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,128,0.05725422170427111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,512,0.005858666780922148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,65536,2.239924536810981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,2560,1.24526490105523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,64,0.0027031112048361036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,2048,0.13712444570329455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,512,0.010621333287821876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,64,0.0053377776510185665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,3584,0.018393778138690524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,8192,0.012454222473833295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,65536,9.892176310221354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,768,0.007361777954631382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,6144,0.14822044637468126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,128,0.012975999878512489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,256,0.005111111121045219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,768,0.006873777757088344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,32,0.00258933338854048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,16384,0.10249688890245225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,768,0.0063288890653186375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,7168,0.02591733303334978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,1024,0.09247644742329915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,65536,0.8933413823445638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,3072,0.027279999521043565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,3584,0.008023999631404877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,51200,0.05776888794369168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,10240,0.00629333323902554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,65536,0.18770133124457467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,16384,0.10497244199117024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,2048,0.015421332584487068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,32,0.03619911273320516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,7168,0.033999111917283796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,768,0.0057591112951437635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,128,0.008157333566082848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,65536,0.0651600029733446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,1536,0.007255111303594377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,7168,0.024413333998786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,1024,0.007778666913509369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,5120,0.01183999992079205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,16384,0.06089244286219279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,256,0.004322666674852371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,7168,1.615398195054796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,1536,0.04217155443297493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,51200,0.8435253567165799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,10240,0.1294755538304647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,5120,0.2657066716088189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,5120,0.021543110410372417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,16384,0.0195795562532213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,4096,0.011776888536082374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,4096,0.06808977656894259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,1024,0.02199911077817281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,12288,0.032511999209721885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,768,0.003935111065705617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,6144,0.20936266581217447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,7168,0.28417245546976727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,3072,0.15570310751597086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,7168,0.006615110983451207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,8192,0.11851822005377875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,6144,0.018834667073355783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,8192,0.4692844284905328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,3072,0.011697777443461947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,32,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,5120,0.3586897850036621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,6144,0.05116800136036343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,2560,0.020922667450375024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,5120,0.022686221533351477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,6144,0.01825777855184343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,65536,0.11787554952833389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,512,0.05459288756052653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,16384,0.10529510842429267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,1536,0.017287110288937885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,768,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,768,0.006454222318198945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,10240,0.02776799930466546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,4096,0.07967199881871541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,128,0.022872888379626807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,128,0.00979466653532452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,768,0.00508977762526936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,768,0.007954667011896769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,8192,0.02927288744184706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,7168,0.015376889043384127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,5120,0.01585155559910668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,4096,0.11675555176205105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,3072,0.1351662211947971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,2048,0.017476444443066914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,51200,2.63411373562283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,5120,0.31044088469611275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,12288,0.01605599953068627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,2048,0.010299555957317352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,512,0.07673866881264581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,3072,0.024015110399987962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,1536,0.08293155829111735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,3072,0.006615110983451207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,3072,0.19150311417049834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,1536,0.023507555325826008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,768,0.019093333019150626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,768,0.008524444368150499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,256,0.02055111030737559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,128,0.15197422769334581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,7168,1.0537972980075412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,128,0.010592888626787396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,5120,0.016002666619088914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,128,0.003020444470975134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,6144,0.031452443864610456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,16384,4.891035715738933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,7168,0.2794524563683404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,7168,0.1501235564549764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,5120,1.5125279956393773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,10240,0.2649466726515028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,65536,1.6285048590766058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,16384,0.06400711006588407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,768,0.0273324449857076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,128,0.012843555874294706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,65536,0.1321671141518487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,2048,0.02011022302839491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,12288,0.3632524543338352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,3584,0.015377778146002026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,3072,0.007637333538797166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,2048,0.10337422291437785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,51200,0.21451111634572348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,1024,0.011368888947698804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,512,0.004864888886610667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,12288,0.021337777376174927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,10240,0.06985066996680366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,65536,0.7084764374627007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,768,0.016978666186332703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,4096,0.008581333690219456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,1024,0.019134221805466544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,3584,0.020601777566803824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,10240,0.023563555545277063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,65536,0.018985778093338013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,5120,0.006219555520349079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,4096,0.013167999684810638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,2560,0.11704088581932916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,2560,0.1526257726881239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,7168,0.03610666592915853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,512,0.004624888714816836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,51200,3.215120103624132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,256,0.007595555649863348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,3584,0.009866666462686326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,51200,0.22532354460822213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,768,0.005843555761708154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,12288,0.025250666671329077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,5120,0.02015822297996945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,2048,0.008600888980759515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,16384,0.023530667026837666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,3072,0.011140444212489657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,64,0.003112888791494899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,16384,0.013815999858909182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,5120,0.06043644746144613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,16384,0.03287200133005778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,65536,0.06536799669265747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,2048,0.01219555570019616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,128,0.002814222127199173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,256,0.0032151111712058387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,10240,0.02920177910063002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,7168,0.016384000579516094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,8192,0.02971733278698391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,32,0.0026088888860411118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,512,0.0032799999333090256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,6144,0.02180266711446974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,3072,0.018977777825461496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,10240,0.06769777668846978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,32,0.007265778051482306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,1024,0.010704889065689512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,2560,0.06373955806096394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,1024,0.016549333930015564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,512,0.009775111244784461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,5120,0.19176355997721353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,3072,0.2283031145731608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,7168,0.3054737779829237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,4096,0.09285155269834731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,512,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,3072,0.04892888996336195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,32,0.005223999834722943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,2048,0.048058665461010404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,32,0.005233777893914117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,65536,0.06390666961669922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,128,0.0028159999185138275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,12288,60.62786187065972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,1536,0.011349333657158745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,1536,0.11731733216179742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,10240,0.3704480065239801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,2048,0.016270221935378183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,2048,0.04584355486763848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,512,0.006680888848172293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,3584,0.01819022165404426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,3584,0.008674666285514832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,4096,0.008974221845467886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,1024,0.008066666622956594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,2560,0.06865600082609388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,1024,0.005581333405441708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,12288,0.04313155677583483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,65536,1.7202604081895616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,2560,0.008620444271299574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,256,0.004433777597215441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,1536,0.006508444332414203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,12288,0.022608889473809138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,7168,0.5987777709960938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,1536,0.05130399929152595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,8192,0.014408000641398959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,16384,0.015117333994971381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,5120,0.02683466672897339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,5120,0.011301333705584208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,8192,0.06236355834537082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,4096,0.10908177826139663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,768,0.0036319999231232535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,1024,0.015431111057599386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,5120,0.10346577564875285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,12288,0.0669599970181783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,16384,0.048728889889187284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,12288,0.08879999981986152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,8192,0.022841778066423204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,51200,0.07044977611965604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,12288,0.09461866484747992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,16384,0.04108000132772658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,128,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,8192,0.01796266602145301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,768,0.004372444417741564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,1536,0.09175911214616563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,4096,0.027440889014138117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,6144,0.010055999788973067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,512,0.033350222640567355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,512,0.0086986662613021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,16384,0.04831288920508491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,8192,0.015071110592948066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,10240,0.006308444258239534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,6144,0.08348088794284397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,128,0.017455110947291057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,1536,0.00814400033818351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,16384,0.0599857767422994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,7168,0.014344889256689282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,7168,0.6892711321512858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,1536,0.005029333134492238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,5120,0.013217777841620974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,8192,2.6843556298149953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,768,0.015136889285511441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,3584,2.7592612372504344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,5120,0.014153778553009033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,2560,0.008739555875460306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,128,0.003942222230964237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,10240,2.9835963779025607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,6144,0.015357333752844067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,16384,0.2593262195587158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,1024,0.0038622220357259116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,8192,0.28634222348531085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,512,0.0038657776183552216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,1024,0.04197688897450765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,5120,0.15608710712856716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,1024,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,65536,0.14869156148698595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,10240,0.275582234064738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,2560,0.019633778267436557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,512,0.047015110651652016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,3584,0.20612800121307373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,1024,0.03898222247759501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,10240,0.00646311127477222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,1536,0.2673715485466851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,12288,0.028759999407662287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,128,0.006565333240562015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,4096,0.028561777538723413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,4096,1.0258684158325195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,65536,0.08908710877100627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,10240,0.05210222138298882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,3072,0.010816888676749336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,256,0.010625777973069085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,128,0.007105777661005656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,512,0.011614222493436603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,6144,0.27984534369574654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,16384,0.12612533569335938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,16384,0.02418488926357693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,512,0.0053297777970631914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,2560,0.01886399918132358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,32,0.26622221204969615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,4096,0.06315110789404975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,3584,1.0529261695014107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,8192,0.015091554986106025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,10240,0.24316623475816515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,16384,0.6914800008138021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,3072,0.015111999379263984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,8192,0.07760444614622328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,4096,0.010254222485754225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,10240,0.014369777507252164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,7168,0.008825778133339351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,65536,0.9476577970716688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,51200,0.06849244568083021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,512,0.13999911149342856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,128,0.0032417778339650896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,32,0.003485333174467087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,7168,0.25272266070048016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,32,0.01611199975013733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,512,0.003302222324742211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,512,0.060991108417510986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,2048,0.06403911113739014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,2048,0.010679999987284342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,512,0.008584889272848764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,512,0.005345777918895085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,32,0.0027217778066794076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,2560,0.23523823420206705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,3584,0.009732444253232744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,1024,0.032425777779685125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,128,0.008194666769769456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,5120,0.03662577933735318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,512,0.012655110822783576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,256,0.012331555287043253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,16384,0.049344887336095176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,3072,0.013245333400037555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,4096,0.14432266023423937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,5120,0.1728008853064643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,768,0.012239110966523489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,7168,0.01738755570517646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,16384,0.008101333346631791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,512,0.011502222054534487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,10240,0.014304889572991265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,8192,0.024141333169407312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,1536,0.013006221916940479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,64,0.007791111038790808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,8192,0.10363555616802639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,2048,0.00868622213602066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,1536,0.01905066602759891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,16384,0.04314755400021871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,65536,0.010787555740939246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,10240,0.06789688931571113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,12288,0.016198222835858662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,3072,0.006704888823959563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,10240,0.0177484436167611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,512,0.009150221943855286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,2560,0.006980444822046492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,768,0.05027022295527988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,768,0.1441377798716227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,768,0.004056000047259861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,8192,0.014428445034556918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,10240,0.21326843897501627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,10240,0.25421423382229275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,64,0.03666222095489502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,1024,0.08119110928641425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,256,0.038167999850379095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,7168,0.023344000180562336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,51200,1.4465920130411785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,4096,0.022478222846984863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,16384,0.5966915554470485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,65536,17.83593580457899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,65536,0.6752755377027723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,4096,0.007920000288221572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,1024,0.010363555616802638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,512,0.3008942339155409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,3584,0.009572444690598382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,3072,0.0284186667866177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,64,0.004606222112973531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,4096,0.1424808899561564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,768,0.05922222137451172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,256,0.01054488867521286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,128,0.006586666736337874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,1536,0.01924000018172794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,1024,0.004844444493452708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,2048,0.09074933661354913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,7168,0.07060888740751478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,32,0.002785777880085839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,10240,0.00925155563486947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,8192,0.278564453125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,1536,0.030223998758527968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,10240,0.4236302110883925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,256,0.0058808889653947614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,2560,0.00870488915178511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,8192,0.589985794491238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,2560,0.017087111870447796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,2560,0.017086222767829895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,6144,0.022060443957646687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,1024,0.017510222064124215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,8192,0.05966133541531033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,2560,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,6144,0.01566133399804433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,4096,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,65536,1.0790987014770508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,3584,0.13842666149139404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,65536,0.19356089168124727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,768,0.006595555692911148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,16384,0.4892115592956543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,512,0.005702222386995952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,7168,0.15396000279320612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,8192,0.028769777880774602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,4096,0.01352533366945055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,6144,0.015588444140222339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,3584,0.006677333265542984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,5120,0.04696444339222378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,256,0.006471111128727595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,3584,0.008386666576067606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,51200,0.29232798682318795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,7168,0.34861156675550675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,8192,0.22604444291856554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,10240,0.1514471107059055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,3072,0.011807111402352652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,1536,0.04010577665434943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,1536,0.12611999776628283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,12288,0.08488621976640488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,7168,0.012694222231705984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,64,0.03339733348952399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,51200,0.07003377543555366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,768,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,256,0.009309333231714036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,16384,0.019170666734377544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,3072,0.05456088980038961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,5120,0.012373333175977072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,128,0.020799999435742695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,32,0.00963466697269016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,10240,0.10710844728681777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,1024,0.042936000559065074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,7168,0.011186666786670685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,256,0.003083555648724238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,1024,0.04004977809058295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,512,0.0034302222645945023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,10240,0.014477334088749357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,256,0.004775111046102312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,1536,0.041879110866122775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,1024,0.0732497771581014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,1024,0.0876435571246677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,1024,0.029005332125557795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,12288,0.035056889057159424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,1024,0.014898666077189975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,1536,0.004552000098758274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,7168,0.058911111619737416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,64,0.007173333730962541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,12288,0.040129777457978986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,2560,0.07433778047561646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,16384,0.0843848917219374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,5120,0.012698666916953193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,2560,0.2816150983174642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,7168,0.08084799846013387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,768,0.0036115555299652945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,1536,0.009507555928495195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,12288,0.1908408933215671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,16384,0.010416000253624385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,32,0.0026293332791990707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,2560,0.007401777638329401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,12288,0.041570666763517596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,51200,0.17131466335720488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,2048,0.10832533571455215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,1024,0.017024889588356018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,8192,0.22520711686876085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,5120,0.2134231064054701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,32,0.03813955518934462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,128,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,1536,0.006261333409282897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,4096,0.13587289386325413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,65536,2.996890597873264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,65536,5.396313985188802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,5120,0.03918222255176968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,4096,0.02029333346419864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,65536,0.016321778297424316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,3584,0.008576889004972246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,3072,0.011301333705584208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,512,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,5120,0.0745848880873786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,7168,0.035537779331207275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,1024,0.4941377639770508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,128,0.004190222256713443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,2048,0.07240088780721028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,10240,0.03602577911482917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,4096,0.030001779397328694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,2560,0.11332621839311387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,65536,0.30261156294080943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,51200,0.11558755238850911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,3584,0.1617342233657837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,7168,0.09535377555423313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,4096,0.015670221712854173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,256,0.016265778077973258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,3584,0.054915553993648954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,51200,2.488830142550998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,51200,13.194632636176216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,2048,0.009732444253232744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,6144,0.01260444439119763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,3072,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,256,0.0210817770825492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,512,0.023606222536828782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,256,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,12288,0.05609689156214396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,2048,0.007436444362004598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,10240,0.06892889075809054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,2048,0.14314044846428767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,5120,0.3465840021769206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,12288,0.12486310799916585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,1536,0.01458755632241567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,32,0.017785777648289997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,3072,0.1191271146138509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,1024,0.0041999999019834726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,2048,0.01290933373901579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,12288,0.19742222627003989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,4096,0.29995467927720815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,3072,0.008264888491895463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,1024,0.012717333104875354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,2048,0.02420622275935279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,128,0.03587733374701606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,1024,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,1536,0.006264888991912206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,2560,0.014581332604090372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,10240,0.08955999877717759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,32,0.006100444330109491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,8192,0.07069777780108981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,6144,0.057894223266177706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,8192,0.029233776860766943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,1024,0.0064399999876817065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,12288,0.11064710881974961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,8192,0.012157333393891653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,1024,0.029356443219714697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,2048,0.019336889187494915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,1536,0.011969777444998423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,1024,0.4137849013010661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,768,0.017981333865059745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,2560,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,2560,0.005640000104904175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,2560,0.03373688790533278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,4096,0.025759999950726826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,65536,0.5406213336520725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,256,0.010168000227875179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,1024,0.016082665986484952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,16384,0.25456354353162974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,2560,0.00832622249921163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,16384,5.246799045138888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,6144,1.4394711388481989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,10240,0.14506133397420248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,65536,1.495939572652181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,10240,0.11070044173134698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,128,0.015094222293959724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,16384,0.021055110626750525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,12288,0.05848711066775852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,51200,0.09607733620537652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,512,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,32,0.010293333066834344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,32,0.00777511133088006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,128,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,768,0.003555555724435382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,8192,0.1747359964582655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,32,0.01587377819750044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,12288,0.03309511144955953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,256,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,32,0.0029031110720502008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,2560,0.02473066747188568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,65536,0.7898337576124402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,64,0.0026008888251251644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,6144,0.01182044463025199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,1536,0.5495262145996094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,12288,0.609283553229438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,12288,0.017243555850452848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,3584,0.008904889225959778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,2048,0.01090400003724628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,2048,0.010631110933091907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,5120,0.01347733371787601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,2560,0.007730666961934831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,10240,0.048727999130884804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,1024,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,128,0.004118222329351637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,3584,0.017224889662530687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,51200,11.036732143825956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,4096,0.3409599992964003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,4096,0.05084622237417433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,16384,0.04294755392604404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,256,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,16384,0.011672000090281168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,6144,0.0460382236374749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,3072,0.008380444513426887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,1536,0.007061333292060428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,4096,0.006855999843941794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,2048,0.005621333503060871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,1024,0.03635288940535651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,2048,0.00980000032318963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,5120,0.09172177977032131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,5120,1.1448870764838326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,3584,0.05656533108817207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,128,0.0029333333174387612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,51200,2.4695627424452042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,256,0.0047235555118984645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,1536,0.02273955610063341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,12288,0.16399999459584555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,2560,0.014583110809326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,16384,0.3275715510050456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,5120,0.31507910622490776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,7168,0.006816888848940532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,10240,0.0854640007019043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,5120,0.135752002398173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,4096,0.010379555324713389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,3584,0.14443555143144396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,1536,0.007205333146784041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,51200,0.5036160151163737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,768,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,2048,0.007708444363541073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,7168,0.07816000117195977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,2048,0.008074666890833113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,65536,1.5247484842936199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,12288,0.03142222099834018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,51200,0.1092746655146281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,3072,0.08332710795932345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,16384,0.018910222583346896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,51200,3.8631697760687937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,256,0.10314310921563043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,32,0.009301332963837517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,7168,0.06293866369459364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,512,0.011508444117175208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,1536,0.1541759967803955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,1536,0.11958577897813584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,256,0.03975555631849501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,7168,0.2731528811984592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,1024,0.04679288797908359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,1536,0.008355555435021719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,51200,0.08115288946363661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,7168,2.068858676486545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,51200,31.686764187282986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,768,0.00832799987660514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,16384,0.0756959981388516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,1024,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,256,0.014383999837769402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,128,0.010844444235165914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,768,0.006144888699054718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,65536,0.36248265372382266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,5120,3.8479415045844187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,8192,0.018869333797030978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,2560,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,12288,0.022615111536449854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,1024,0.0075031113293435835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,512,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,6144,0.010474666953086853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,6144,0.02515911062558492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,10240,0.31301866637335884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,128,0.08475022183524238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,16384,0.13842666149139404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,512,0.07423200209935506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,1024,0.30465155177646214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,65536,2.3618026309543185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,1024,0.012155555188655853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,1024,0.0181031111213896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,3584,0.010101333260536194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,64,0.0031848889258172777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,256,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,5120,1.169342252943251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,6144,0.05016977919472588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,32,0.0032053333189752367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,1536,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,512,0.008982222113344405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,512,0.007028444773621029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,65536,0.4812657568189833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,32,0.0033662221911880705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,2048,0.007052444749408298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,256,0.22216888268788657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,256,0.039650665389166936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,5120,0.13709421952565512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,3072,0.0188755558596717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,12288,0.23638576931423613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,512,0.0033813334173626373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,5120,2.3292115529378257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,12288,0.07912977536519368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,6144,0.011487111449241638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,64,0.0027111110587914786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,3584,0.06736799743440416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,6144,0.06891466511620416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,4096,0.06631555822160509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,2048,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,64,0.013372444444232516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,16384,1.694234636094835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,16384,0.02532177832391527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,3072,0.018292443619834054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,3584,0.08159199688169691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,512,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,768,0.0123857781291008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,768,0.0037751110891501107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,16384,0.031066666046778362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,51200,0.13117867045932347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,51200,0.877512878841824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,1536,0.00997155573632982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,16384,0.021503110726674397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,1536,0.028747555282380845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,3072,0.008055110772450766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,65536,1.5961244371202257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,5120,0.010997333460383944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,3584,0.0833466649055481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,1536,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,5120,0.022316444251272414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,256,0.007643555601437886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,12288,0.016795555750528973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,8192,0.21578755643632677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,2048,0.010828444527255164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,3584,0.1890853378507826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,5120,0.025877333349651758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,512,0.009716444545321995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,3584,0.01367377738157908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,4096,0.015071110592948066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,16384,0.07392533620198567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,4096,0.013821333646774292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,12288,0.8865502145555285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,12288,0.007034666836261749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,6144,0.19804888301425508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,64,0.008453333543406593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,7168,0.014493332968817817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,3584,0.2096640004052056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,2048,0.01441422270403968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,12288,0.36199378967285156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,1024,0.03687199950218201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,768,0.003996444245179494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,512,0.0058551111982928375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,3072,0.0491848886013031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,8192,0.006689777804745569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,51200,0.10627644591861302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,1024,0.004206222378545337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,128,0.008336889247099558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,3584,0.01829333272245195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,6144,0.015592000550693937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,12288,0.03670310974121094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,8192,0.007977777885066139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,1536,0.00489155575633049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,3584,0.009810667071077559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,3584,0.0655848913722568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,51200,0.06462044186062284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,256,0.0303111109468672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,7168,0.01713244451416863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,3584,0.24338934156629774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,256,0.00452888881166776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,4096,0.11931822035047744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,3584,0.14747644795311823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,5120,0.09655821985668606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,128,0.005772444523043103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,8192,0.14731644259558782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,1024,0.009115555220180089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,768,0.005964444329341252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,768,0.008788444101810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,1536,0.006165333506133821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,128,0.008792888787057664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,1024,0.01621955633163452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,2048,0.039361778232786394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,2048,0.011362666885058085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,10240,0.06367644336488512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,3584,0.031038221385743883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,10240,0.0895306666692098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,65536,0.17525955041249594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,256,0.0028248888750871024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,64,0.05566222137875027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,3072,0.029402666621738013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,51200,0.3044453461964925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,65536,0.0754657785097758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,2048,0.054245332876841225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,7168,0.01737155516942342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,1536,0.006315555423498154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,4096,0.030943112240897283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,5120,0.04984711276160347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,768,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,2560,0.012864888542228274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,2048,0.011753777662913004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,3072,0.3521582285563151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,256,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,512,0.0034311111602518293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,32,0.010768888725174798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,65536,0.06484178039762709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,16384,0.13292977544996473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,10240,0.03653066688113742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,128,0.004912000149488449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,16384,1.4140915340847438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,7168,0.03234488765398661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,2048,0.008730666504965888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,1536,0.18014132976531982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,64,0.00943911075592041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,512,0.02270577847957611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,6144,0.008621333373917473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,51200,0.18374844392140707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,8192,5.909687889946832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,12288,0.08624977535671657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,768,0.010779555473062726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,1024,0.012453333371215396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,12288,4.1549224853515625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,768,0.010692444112565784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,16384,0.05277511146333483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,16384,0.23718399471706816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,512,0.0034337778472238113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,3072,0.010647999743620554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,4096,0.18925155533684623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,256,0.036088890499538846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,768,0.017549332645204332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,12288,0.2657528983222114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,65536,0.009638222555319468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,3584,0.17144889301723906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,512,0.005035555611054103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,256,0.011600889265537262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,256,0.0034942221310403612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,16384,0.44771819644504124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,16384,0.03449777762095133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,512,0.0063386667105886675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,256,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,8192,0.04004177782270644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,768,0.2136533392800225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,64,0.0027146666414207886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,768,0.02755644420782725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,10240,0.6236329078674316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,51200,0.08578844202889337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,2560,0.013208000196350945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,7168,0.015048889650238885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,8192,0.01643022232585483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,10240,0.09279733233981663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,6144,0.09520622094472249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,128,0.016539555456903245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,51200,0.2688586711883545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,128,0.01129244433508979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,10240,0.023408888114823237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,51200,0.00798044436507755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,3072,0.009180444810125563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,65536,0.14572800530327693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,3584,0.014128888646761576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,256,0.004652444273233414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,32,0.005371555685997009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,128,0.010467555787828235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,3584,0.00870577742656072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,7168,0.08085599872801039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,3072,0.028002666102515325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,10240,0.013953778478834363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,1536,0.06410844458474053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,1024,0.0058808889653947614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,3584,0.00792711145348019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,256,0.004569777597983678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,2048,0.022500443789694045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,7168,0.19687555895911324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,32,0.21021066771613228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,2560,0.007556444240940942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,7168,0.1327644454108344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,1024,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,16384,0.024680889315075342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,32,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,2048,0.009913778139485253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,3072,0.07766933573616876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,65536,0.0690844456354777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,1024,0.005384888913896348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,8192,0.02197066611713833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,3584,0.01607733302646213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,1536,0.01586577792962392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,12288,0.06588977575302124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,2048,0.015051555302408008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,5120,0.01981511049800449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,5120,0.26070933871799046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,128,0.017850667238235474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,5120,0.17757511138916016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,7168,0.023318222827381555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,12288,0.05018933282958137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,4096,0.010089777410030365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,10240,0.09789955615997314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,1024,0.008192888564533657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,2048,0.01051288925939136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,65536,0.13220177756415472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,5120,0.02198311189810435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,7168,0.27596622043185765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,2048,0.01978400018480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,64,0.0028853333658642242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,512,0.013100444442696042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,12288,0.00738933351304796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,256,0.004247111164861255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,768,0.016778666112158034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,2048,0.043234666188557945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,768,0.020153777466879953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,4096,0.01313066648112403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,6144,0.08510133292939927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,16384,0.3881244394514296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,7168,0.025220443805058796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,7168,0.019128888845443726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,256,0.018954666124449838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,5120,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,12288,1.2592133416069877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,256,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,768,0.007479111353556315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,7168,0.021503110726674397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,65536,2.5365111033121743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,6144,0.17378577921125624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,51200,0.06673510869344075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,64,0.005480888817045424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,4096,0.010977778169843884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,4096,0.07899377743403117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,32,0.00424888895617591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,1024,0.01238044434123569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,32,0.009172444542249044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,65536,0.2846897708045112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,1536,0.01670222149954902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,5120,0.033484445677863225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,1024,0.011502222054534487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,4096,0.01941688855489095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,4096,0.02241155505180359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,1024,0.026597332623269822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,2560,0.012006222373909421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,65536,0.06910666492250231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,16384,0.07557688819037543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,8192,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,6144,0.025622222158643935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,51200,1.371646245320638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,16384,0.01536977787812551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,512,0.0068915556702348925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,2560,0.010433778166770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,2048,0.007879111501905654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,16384,0.06452266375223796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,32,0.02184711065557268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,2048,0.032446222172843084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,8192,0.020679111282030743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,1536,0.0192586663696501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,1024,0.00620977787507905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,3584,0.02548622257179684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,2048,0.07530933618545532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,16384,0.1732684506310357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,512,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,512,0.01401600076092614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,768,0.004201777693298128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,512,0.01574844453069899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,7168,0.011052444577217102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,51200,0.6277350849575466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,5120,0.0319164428446028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,1536,0.008656889200210571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,1536,0.005864000154866113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,16384,0.06192533175150553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,65536,0.06879733006159465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,12288,0.0440008905198839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,2048,0.013225778109497495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,256,0.003020444470975134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,512,0.10283466842439438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,1024,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,3584,0.6655226813422309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,10240,0.0729315545823839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,4096,0.15371821986304388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,256,0.004936000125275718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,65536,1.9850230746799047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,1536,0.02204622162712945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,65536,0.2165395551257663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,2048,0.04830311073197258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,2048,0.02605155607064565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,51200,4.254358079698351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,2560,0.15697067313724095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,768,0.005670222143332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,768,0.009062222308582729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,7168,0.0070426662762959795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,1536,0.0564240018526713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,4096,1.076672871907552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,2048,0.07260711325539483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,512,0.007502222226725684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,4096,0.01258133351802826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,5120,0.040223111708958946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,12288,0.04350044329961141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,6144,0.05316533313857185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,256,0.025955556167496577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,3072,0.061095111899905734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,16384,0.01867911054028405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,6144,1.4070844650268555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,4096,0.008540444076061249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,4096,0.7209502326117622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,4096,0.02957333458794488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,6144,0.10070488850275676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,4096,0.019311111834314134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,12288,0.09661155276828343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,64,0.002607999990383784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,2560,0.04100088940726386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,1024,0.010144888526863521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,2048,0.02906044324239095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,32,0.0034106667670938703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,1024,0.1357777780956692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,3584,0.01829155617290073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,6144,0.020276443825827706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,2048,0.026661333110597398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,1536,0.012345777617560493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,12288,0.033815112378862165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,65536,0.13435288270314535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,51200,0.06520089175966051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,3072,0.12883910867902967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,16384,0.44715197881062824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,512,0.003338666632771492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,10240,0.017654221918847825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,1536,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,51200,4.0983662075466585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,768,0.08913599782519871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,768,0.01051111105415556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,2048,0.011699555648697747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,2048,0.00609422226746877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,16384,0.01570133368174235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,10240,0.035359111097123884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,7168,0.008198222352398766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,512,0.004870222260554631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,51200,0.07465333408779569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,1536,0.15326844321356878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,51200,0.11893333329094781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,4096,0.1472782161500719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,12288,0.039814223845799766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,3584,0.02568622264597151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,1536,0.016164445214801364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,65536,3.567607879638672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,51200,0.029567999972237483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,768,0.006783999916579988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,32,0.009533333281675974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,6144,0.2520577907562256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,6144,0.46761422687106663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,8192,0.06170044342676798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,3072,0.009486222432719337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,51200,0.01847644481394026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,3584,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,12288,0.03764977720048692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,51200,0.13703733020358616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,5120,0.03197155396143595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,64,0.005164444446563721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,1536,0.005384888913896348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,2048,0.009495110975371467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,512,0.0038319999972979226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,3072,0.009591110878520543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,1536,0.02975644336806403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,256,0.0047306666771570844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,3072,0.030284444491068523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,6144,0.012375999655988483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,4096,0.0630319979455736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,4096,0.0794995559586419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,768,0.014476444986131458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,1536,0.0046844445168972015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,1536,0.008029333419269985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,12288,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,16384,0.07972000042597453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,51200,0.05455999904208713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,1536,0.01649599936273363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,1024,0.007154666715198093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,2560,0.011663999822404651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,1024,0.00962755580743154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,8192,0.031925333870781794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,3072,1.355184025234646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,65536,2.4587164984809027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,2048,0.007808888951937358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,5120,0.01349955548842748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,4096,0.06810400221082899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,4096,0.01367911116944419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,7168,0.5004266632927789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,6144,2.6724035474989147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,512,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,4096,0.02010933392577701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,65536,0.0594924423429701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,32,0.22204089164733887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,7168,0.010878221856223213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,10240,0.037674665451049805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,512,0.11011555459764268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,512,0.19066666232215035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,512,0.005207111024194294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,51200,0.029306666718588933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,512,0.03420533405409919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,4096,0.02894222074084812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,65536,0.57059997982449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,12288,0.015763555963834126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,2560,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,2560,0.022687999738587275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,51200,0.6068319744533962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,1024,0.011209777659840055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,768,0.017650667164060805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,1536,0.1810844474368625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,12288,0.07292533583111234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,65536,0.9926417668660482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,2560,0.008489777644475302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,64,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,2560,0.024480889240900677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,768,0.02829244401719835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,16384,0.0763840013080173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,1024,0.005405333307054307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,5120,0.010533333652549319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,10240,0.05150399936570061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,256,0.03762044509251913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,1536,0.39353511068556046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,5120,0.020833777056799993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,6144,0.008786666724416945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,51200,4.458267635769315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,128,0.013515555196338229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,2048,0.1405173275205824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,64,0.017326222525702584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,2560,0.00850844466023975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,2048,0.030943999687830608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,51200,0.8137857649061414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,128,0.010803555448849997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,16384,1.9534515804714625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,10240,0.06291821930143568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,8192,0.016473778420024447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,4096,0.40211288134257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,512,0.04551022251447042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,4096,0.013760888742076026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,6144,0.019492444064882066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,256,0.06746844450632732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,5120,0.02368444369898902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,128,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,16384,0.01959911154376136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,8192,0.47562668058607316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,51200,0.37180089950561523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,1024,0.048325333330366343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,3584,0.00666933341158761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,8192,1.1738711463080513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,1024,0.01092266705301073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,2048,0.15941511260138616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,256,0.012316444681750404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,128,0.02844888965288798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,256,0.1164097785949707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,2048,0.022602667411168415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,768,0.005719999886221356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,2560,0.17152888245052764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,6144,0.06667288806703356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,768,0.05361510978804695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,32,0.024352888266245525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,65536,0.2927190992567274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,10240,0.08081066608428955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,1024,0.004063111212518481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,7168,0.024077332682079736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,65536,0.41580266422695583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,768,0.028304888142479792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,5120,0.4022471110026042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,5120,0.03584622343381246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,4096,2.727842754787869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,10240,0.027112000518374976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,768,0.023432888918452795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,1024,0.015776889191733468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,7168,0.024529778295093115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,16384,0.03763644562827216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,16384,6.605291578504775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,64,0.008845333423879411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,16384,0.4255297713809543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,51200,1.263554679022895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,10240,0.017779555585649278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,3072,0.010826666322019366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,512,0.017624000708262127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,1536,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,64,0.014093332820468478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,512,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,2048,0.0052968888647026485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,256,0.006617777877383762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,1536,0.1254391140407986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,256,0.011556444068749746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,128,0.008823999928103553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,32,0.0028391112056043413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,8192,0.028548445966508653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,7168,2.798087014092339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,2048,0.1097066667344835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,256,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,256,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,32,0.0038337777886125776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,7168,0.01163733336660597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,128,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,51200,0.033999111917283796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,32,0.0026808888134029177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,12288,1.3550506167941625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,4096,0.07285600238376193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,32,0.006088888893524806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,2560,0.0379671123292711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,768,0.010625777973069085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,4096,0.08837688631481594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,2560,0.013696889082590738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,65536,0.06222133504019844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,8192,0.08868533372879028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,32,0.031326221095191106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,3072,0.027833776341544256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,2048,0.07664089070426093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,16384,0.09004177649815877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,128,0.1366879940032959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,8192,0.06386755572424994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,5120,0.33572178416781956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,51200,2.54620361328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,4096,0.011748444702890186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,32,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,7168,0.025119110941886902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,3072,0.0090524446633127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,7168,0.12250488334231907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,5120,0.35256266593933105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,8192,0.0328800015979343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,256,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,16384,2.7770426008436413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,65536,2.8014924791124134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,2048,0.6152302424112955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,64,0.006503110958470239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,6144,0.22814133432176378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,5120,0.03289244572321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,32,0.003022222262289789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,7168,0.007248889240953658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,3584,0.059189332856072314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,16384,0.023120888405376013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,2048,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,3072,0.01261333293384976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,1024,0.004358222087224324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,512,0.0034773333205117118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,16384,0.036293334431118436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,65536,1.4659173753526475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,2048,0.06646400027804904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,2048,0.008351999852392409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,64,0.02500533395343357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,10240,0.047192000680499606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,1536,0.012874667015340594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,3584,0.04758488800790575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,3584,0.012493333054913415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,6144,0.01460800071557363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,128,0.004056888735956616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,3584,0.02132266594303979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,10240,0.3267831007639567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,256,0.004774222357405556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,5120,0.08571110831366645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,4096,0.008938666847017076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,3584,0.007167110840479533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,51200,0.010963555839326648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,16384,0.026690666874249775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,1536,0.004785777793990241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,1536,0.14444533983866373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,512,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,3072,0.012287110918098025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,8192,0.04702933298216926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,512,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,6144,0.01015644437736935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,1024,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,256,0.005048888838953442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,65536,0.918066660563151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,2560,0.009961778091059791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,51200,0.6052604251437718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,512,0.07931822538375854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,2048,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,51200,2.6093741522894964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,4096,0.11771999465094672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,7168,0.017488888568348356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,2048,0.01567911108334859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,3072,0.04698488778538174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,512,0.03559911251068115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,512,0.004263111286693149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,8192,0.017649778061442904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,256,0.0070062221752272705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,8192,0.014072888427310519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,768,0.008829333715968663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,1536,0.006350222147173352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,512,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,1024,0.004401777767472797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,32,0.010315555665228102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,65536,0.07993600103590223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,512,0.006132444573773279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,10240,0.22393332587348092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,64,0.002603555512097147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,1536,0.0074142225914531285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,128,0.0027599999060233435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,65536,1.975394566853841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,7168,0.09008000294367473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,12288,0.07832622528076172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,512,0.14339555634392634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,10240,0.020291555258962844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,8192,0.06358310911390516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,10240,0.10149066978030735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,2560,0.011961778004964193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,2048,0.015000889698664347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,10240,0.40544888708326554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,5120,0.07835377587212457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,2048,0.007438222567240397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,16384,0.053156445423762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,32,0.006093333164850871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,4096,0.009148444566461775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,512,0.004439110971159405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,1536,0.050704889827304415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,6144,0.026305778159035578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,10240,0.013776889277829064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,12288,0.006400888992680444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,256,0.013455111119482251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,1024,0.003755555384688907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,768,0.007656888829337225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,1536,0.0328106681505839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,512,0.003685333248641756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,256,0.004332444320122401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,768,0.014849777022997538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,12288,0.021301334102948506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,4096,0.15722400612301296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,16384,0.2546995480855306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,65536,0.014599111345079211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,256,0.008057778080304464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,512,0.005032000028424793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,2560,0.06835466623306274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,1536,0.058548443847232394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,8192,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,1024,0.006139555739031897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,3072,0.006416888700591193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,65536,0.6386817826165093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,64,0.002552888873550627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,16384,0.008895999855465358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,7168,0.015793777174419828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,7168,0.013775111072593264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,1536,2.870544009738498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,256,0.008097777764002482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,256,0.0068959999415609576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,8192,0.0483822226524353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,1536,0.053293334113227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,7168,0.013826666606797112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,12288,0.0273333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,65536,0.4641902181837294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,1536,0.005600888695981767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,3072,0.02206755512290531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,5120,0.04909155435032315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,4096,0.011890666352377998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,512,0.04853866828812493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,4096,0.01718311177359687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,64,0.0025457777082920074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,2560,0.013170666992664337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,1024,0.01014755583471722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,6144,0.06945600112279256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,1536,0.010402667025725046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,2048,0.008887999587588841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,6144,0.11626222398546006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,16384,0.32601245244344074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,6144,0.022030222747060988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,512,0.004420444369316101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,10240,0.18858667214711508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,51200,1.6413422690497503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,4096,0.06442666716045804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,4096,0.019317333896954853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,3072,0.05668977896372477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,8192,0.016881777180565726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,51200,19.123555501302082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,2048,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,65536,0.12157244152492946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,6144,0.01779111060831282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,512,0.018188445104493033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,65536,0.2788471115960015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,256,0.09576621982786392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,5120,0.05536355574925741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,4096,0.01366222235891554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,10240,0.427565336227417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,256,0.004948444250557157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,6144,0.03162844313515557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,65536,0.3537448777092828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,512,0.00943911075592041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,128,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,256,0.005172444300519096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,5120,0.020245333512624104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,768,0.019018666611777413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,51200,0.04758755697144402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,1536,0.0070426662762959795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,1024,0.01130488928821352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,8192,0.05990578068627251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,10240,0.008406221866607666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,8192,0.05832266807556152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,128,0.005937777873542573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,1536,0.010024000373151567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,51200,2.4034862518310547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,5120,0.18388799826304117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,6144,0.013612444202105204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,16384,0.05928444200091892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,256,0.003069333318206999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,2048,0.06538222233454387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,51200,1.1791590584648979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,12288,0.43837867842780215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,6144,0.036912888288497925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,51200,0.058865778976016574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,128,0.06554755237367418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,5120,1.9539430406358507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,12288,0.05103466577000088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,64,0.0035004446076022256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,3072,0.08465777503119574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,51200,3.5630620320638022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,12288,0.9079191419813367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,12288,0.41986844274732804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,768,0.005998222364319696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,128,0.002985777747299936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,1024,0.007344000041484833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,51200,0.34830133120218915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,1536,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,16384,0.029887109994888306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,8192,0.06434755855136447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,64,0.08161777920193143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,4096,0.05196355448828804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,3072,0.04846222201983134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,6144,0.012553777959611682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,3072,0.020078221956888836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,7168,0.052482667896482676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,128,0.05541511045561897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,16384,0.037440889411502414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,7168,0.05381777882575989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,8192,0.23085244496663412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,65536,0.07696266969045003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,768,0.017843556072976854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,7168,0.13809155093299016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,12288,0.020619556307792664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,768,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,2560,0.0509111119641198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,128,0.0063946665161185795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,12288,0.00693866651919153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,5120,0.6660115453932021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,10240,1.737410651312934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,128,0.05008888906902737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,128,0.006245333287451003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,16384,0.030699554416868422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,3584,0.013251555462678274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,768,0.030015110969543457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,1536,0.016001777516471017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,1024,0.06172888808780246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,768,0.009514666265911525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,3584,0.1428577767478095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,1536,0.016202666693263583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,51200,0.19444710678524443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,7168,0.04100622071160211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,16384,0.05537866552670797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,51200,0.10017689069112141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,65536,0.10634666681289673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,16384,0.1318097776836819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,6144,0.0377146667904324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,12288,1.8069032033284504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,4096,0.009970666633711921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,4096,0.034056001239352755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,32,0.00639111093348927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,512,0.049117333359188504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,1024,0.012239110966523489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,16384,0.1952293316523234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,128,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,5120,0.01612977849112617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,51200,0.013791110780504016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,6144,0.30037689208984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,10240,0.06622755527496338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,3072,0.2447120083702935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,3584,0.009282666775915358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,6144,0.08242933617697822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,4096,0.01942755613062117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,8192,0.03196355700492859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,64,0.011534222298198275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,32,0.007357333269384172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,4096,0.27218932575649685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,1024,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,4096,0.013697778185208639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,3584,0.18223999606238472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,2560,0.06257333358128865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,7168,0.01128533316983117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,51200,0.12334133519066705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,16384,0.2285857730441623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,65536,4.446658664279513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,51200,0.13323111004299587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,7168,0.012799999780125089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,256,0.051719999975628324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,10240,0.011128000087208219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,51200,0.10899111297395493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,256,0.008544888761308458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,3584,0.03076710965898302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,3584,0.007716444631417592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,1536,0.04939377639028761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,10240,0.0801875525050693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,5120,0.01689511040846507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,3072,0.0732062194082472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,51200,0.2116675509346856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,4096,0.007914666500356462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,12288,0.1965413358476427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,5120,0.12194932831658258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,32,0.006509333435032103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,16384,0.24952356020609537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,8192,0.015127999915017022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,3584,0.010276444256305695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,16384,1.0731902652316623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,51200,0.1146426730685764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,4096,0.01109688894616233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,256,0.00498844434817632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,8192,0.014615111880832247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,16384,0.12036621570587158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,3072,0.02496622171666887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,10240,0.11621956030527751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,768,0.00371111101574368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,256,0.004423111263248655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,128,0.0052791109515561005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,51200,0.05973777506086561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,4096,0.07515644364886813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,10240,0.025496888491842482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,64,0.005251555393139522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,128,0.004569777597983678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,1024,0.0068959999415609576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,768,0.01844800015290578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,6144,0.08016266425450642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,51200,3.1352231767442493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,3072,0.029264890485339697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,256,0.00647288892004225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,128,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,51200,0.2601004441579183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,51200,0.13689066304100886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,128,0.012446222205956778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,64,0.14868533611297607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,16384,0.12371022171444362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,65536,0.401600890689426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,65536,0.5740995407104492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,2560,0.021659556362364028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,128,0.0028524444335036804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,51200,10.027647230360243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,4096,0.011364444262451597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,512,0.007420444654093848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,51200,0.07397955656051636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,512,0.08911555343204075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,8192,0.02606755495071411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,3072,0.04811822374661764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,3584,0.014887111054526435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,32,0.005197333378924264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,1536,0.027744000156720478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,12288,0.026033777329656813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,6144,0.04308355516857571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,4096,0.10972177982330322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,64,0.0028560000161329904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,2560,0.006676444576846228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,2048,0.10338222318225437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,3072,0.01620444489849938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,6144,0.7550871107313367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,128,0.004990222139490975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,5120,0.01590133375591702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,1536,0.04964088731341892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,512,0.005710222240951326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,3584,0.020473778247833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,4096,0.04347911145952013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,64,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,3584,0.01314133322901196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,8192,0.02835644284884135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,128,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,7168,0.012492444780137805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,256,0.02071644365787506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,256,0.010629333555698395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,4096,0.0693991118007236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,768,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,1536,0.00646577775478363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,512,0.006078222145636876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,16384,0.03849511014090644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,128,0.004026666697528628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,2048,0.5718533198038737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,7168,0.048057778014077075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,4096,0.009848000274764167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,256,0.009950222240553962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,128,0.033831109603246055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,32,0.005920889063013925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,2560,0.24451022677951387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,12288,0.010790222220950656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,2560,0.009570666485362584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,256,0.010489777558379702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,10240,0.03274666600757175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,128,0.0047235555118984645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,2560,0.06503911150826348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,5120,0.15458756022983128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,1536,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,10240,0.018363555272420246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,65536,0.0107004443804423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,128,0.002728888971938027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,10240,0.01690933273898231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,256,0.012746666868527731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,512,0.0037253333462609183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,5120,0.011439111497667102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,8192,0.26848800977071124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,6144,0.008188444707128737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,8192,0.023895111348893907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,768,0.03615999884075589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,128,0.010029333333174387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,6144,0.23272622956169975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,16384,0.04096800088882446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,768,0.046294222275416054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,768,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,3584,0.18345155980851915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,8192,0.44661421245998806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,5120,0.053710222244262695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,1024,0.454845322502984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,4096,0.0137155552705129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,1536,0.04610310991605123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,6144,0.009768889182143742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,768,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,3072,0.018995554910765756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,1536,0.04568355613284641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,768,0.005856888989607493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,10240,0.009892444643709395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,2560,0.011557333171367645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,51200,0.2726808918846978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,16384,0.06893510950936212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,768,0.006068444500366847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,4096,0.20846754974789092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,6144,0.012764444781674279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,8192,0.009776888622177972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,6144,0.08493244647979736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,768,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,1536,0.0064399999876817065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,64,0.010081777969996134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,128,0.008088000118732452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,10240,0.01962933275434706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,64,0.005471999860472149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,65536,0.909728897942437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,768,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,8192,0.042190220620897084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,1536,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,51200,0.060292442639668785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,16384,0.5377422438727485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,3584,0.01297688898113039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,6144,0.16616443792978922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,256,0.027951998843087092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,10240,0.0064239998658498125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,3584,0.006725333217117522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,512,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,16384,0.02388444377316369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,2048,0.006735111276308696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,10240,0.8758293257819282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,16384,0.03071999881002638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,2048,0.005485333502292633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,128,0.002797333316670524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,3072,0.2868817647298177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,64,0.003636444608370463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,51200,0.5012035369873047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,6144,0.04071377714474996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,5120,0.019461333751678467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,8192,0.03014755580160353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,1536,0.011853333148691388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,512,0.020039111375808716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,5120,0.02035288843843672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,2048,0.018495111001862418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,32,0.002985777747299936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,3072,0.010069333016872406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,6144,0.013178666432698568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,4096,0.00978577799267239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,16384,0.17527466350131563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,7168,0.01535555554760827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,10240,0.04677777820163303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,12288,10.532399495442709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,64,0.0038080000215106537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,12288,0.028530667225519817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,128,0.01570933394961887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,12288,0.739159107208252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,64,0.08443733056386311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,16384,0.06129688686794705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,512,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,768,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,16384,0.3004426691267225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,65536,0.02295733326011234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,512,0.005048000150256687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,768,0.004314666820896996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,64,0.011215111447705163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,10240,0.02416177756256527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,256,0.019342222147517733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,6144,0.02985422147644891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,12288,0.0192266669538286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,6144,0.5277342266506618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,3072,0.012832000023788877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,3584,0.05966844161351522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,12288,0.15461688571506077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,16384,0.018022222651375663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,51200,0.014428445034556918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,1536,0.011977777712874942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,1536,0.006673777682913675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,6144,0.0188071115149392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,8192,0.039996445178985596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,128,0.007136888802051544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,64,0.023018666439586218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,12288,0.687786685095893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,5120,0.05788177914089627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,10240,0.015543111496501498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,5120,0.014704000618722705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,512,0.015796444482273526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,8192,0.02549422283967336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,8192,0.06552355819278292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,2048,0.013416888813177744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,768,0.003448888866437806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,512,0.015612444943851896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,8192,0.007739555504586961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,10240,0.026724444495307073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,10240,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,65536,2.7060044606526694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,32,0.006174222048785951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,65536,0.4560142093234592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,32,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,10240,0.05624000231424967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,768,0.003589333345492681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,5120,0.19317955440945098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,3072,0.007227555745177799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,16384,0.01442399952146742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,512,0.011536888778209686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,2048,0.014004444082578024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,256,0.005037333402368757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,16384,0.03341155582004123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,6144,0.02414488957987891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,51200,0.09529778030183579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,32,0.006678222368160884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,3584,0.008819555242856344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,3584,0.017614222235149808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,32,0.0052435555391841466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,64,0.018016889691352844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,8192,0.012646222280131446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,10240,0.3060719966888428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,1024,0.015499555402331881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,2560,0.05630489190419515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,768,0.05019377668698629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,32,0.015799111790127225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,2560,0.016378665963808697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,2048,0.16433955563439265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,1536,0.033349331882264875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,1536,0.006836444553401735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,2560,0.10452355278862847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,12288,0.0602088901731703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,51200,11.261289808485243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,2560,0.2042000028822157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,512,0.009033777647548253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,12288,0.018925334016482036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,7168,0.6177253193325466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,51200,0.2708764341142443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,768,0.004018666843573253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,512,0.0034711110509104202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,1024,0.005763555566469829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,2560,0.011739555332395764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,6144,4.71910137600369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,128,0.006526222245560751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,512,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,2560,0.0397093329164717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,1536,0.009919111099508073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,16384,0.04829333225886027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,768,0.0074106670088238185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,7168,0.03336266676584879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,2560,0.01055555542310079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,8192,0.04019822345839606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,12288,0.7046453158060709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,128,0.1166951126522488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,10240,0.062133332093556724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,1536,0.007137777904669444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,128,0.005715555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,2560,0.012353777885437012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,768,0.007167110840479533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,16384,0.030176000462638006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,3584,0.010466666685210334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,768,0.01019377758105596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,16384,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,1536,0.07159377468956842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,65536,0.2117697795232137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,256,0.0031191110610961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,5120,0.39250577820671934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,16384,0.027629334065649245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,10240,2.412897745768229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,12288,0.18283643987443712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,65536,0.11214399337768555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,512,0.005481777919663324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,3072,0.6199217902289497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,768,0.004233777936961916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,51200,0.14995466338263616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,51200,0.06121333440144857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,256,0.01201600001917945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,4096,0.05484888950983683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,512,0.0038702223036024305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,7168,0.01423555612564087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,8192,0.012916444076432122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,10240,0.021321778496106465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,2560,0.06757244136598375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,1536,0.0069848886794514125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,4096,0.1355244451098972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,65536,0.9875164031982422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,768,0.2100062237845527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,10240,0.019392000304328073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,51200,6.135825686984592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,1024,0.04838844471507602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,1024,0.08780355585945977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,65536,0.09382488992479111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,64,0.006335111127959357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,2048,0.01640533407529195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,1024,0.36489778094821507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,2560,0.03404266635576884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,32,0.007160888777838812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,2048,0.09497600131564671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,512,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,65536,0.15135733286539713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,32,0.0027946666296985415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,128,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,768,0.011408888631396823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,12288,0.006584888945023219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,2560,0.03996977872318692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,12288,0.018787556224399142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,1024,0.005223111146026188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,1024,0.059261335266960986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,768,0.05194488830036587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,512,0.02459555533197191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,12288,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,1536,0.07353688610924615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,5120,0.09886222415500218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,1024,0.006776000062624614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,3072,0.008558221989207799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,256,0.0243493335114585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,10240,0.17463555600908068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,64,0.005039111193683413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,51200,4.714237213134766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,32,0.06538310978147718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,32,0.016904888881577384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,3072,0.4548337724473741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,4096,0.04249866803487142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,512,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,12288,0.026955554882685345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,256,0.012160888976520963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,3584,0.007711999946170383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,2048,0.007730666961934831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,16384,1.001482645670573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,5120,0.22455644607543945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,4096,0.045455111397637256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,7168,0.08944177627563477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,2560,0.1628613339530097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,16384,0.6614080005221897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,2048,0.022441777918073866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,256,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,1024,0.014013333453072442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,768,0.03378488951259189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,12288,0.35780978202819824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,5120,0.13682044876946345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,512,0.004092444562249714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,512,0.004293333325121137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,2560,0.009524444739023844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,8192,0.013373333546850415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,65536,0.18085155222151014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,6144,0.08112266990873548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,8192,0.016460445192125108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,2048,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,6144,0.013973333769374423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,10240,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,5120,0.15955466694302028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,256,0.004907555464241239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,2560,0.20131377379099527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,1024,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,6144,0.08435200320349799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,10240,0.013752889302041797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,8192,0.08043289184570312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,7168,0.028767112228605483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,32,0.00481599983241823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,12288,0.11648178100585938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,1024,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,7168,1.5839938057793512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,5120,0.14044266276889378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,2560,0.03269333309597439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,2048,0.3427768813239203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,16384,0.3521724541982015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,256,0.007817777494589487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,5120,0.034318221939934626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,8192,0.040579554107454084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,2560,0.13371111287011042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,12288,0.41442134645250106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,12288,0.010111110905806223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,1536,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,128,0.008061333662933773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,64,0.03728355632887946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,10240,0.016717332932684157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,1024,0.13372445106506348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,5120,0.1668097840415107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,8192,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,3584,0.023631999890009563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,2048,0.01884711119863722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,10240,0.11096177498499553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,7168,0.043319112724728055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,256,0.06321333514319526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,10240,0.07540622022416857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,3584,0.012078222301271228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,64,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,7168,0.013891556196742587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,2048,0.005947555518812603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,8192,0.016199111938476562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,1536,0.06020444631576538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,2048,0.03630400035116408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,6144,0.02509333358870612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,12288,0.10214222139782375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,1024,0.02401155564520094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,4096,0.015101333459218344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,65536,0.13752978377872044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,16384,0.0236791107389662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,512,0.0034355556385384668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,7168,0.014236445228258768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,128,0.0026613333159022858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,7168,0.13926045099894205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,10240,0.025265778104464214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,128,0.003999999827808804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,512,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,2560,0.009478222164842818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,12288,0.01815555493036906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,10240,0.011029333704047732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,4096,0.0058328890138202244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,1024,0.24720356199476454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,7168,0.028885334730148315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,512,0.016511999898486666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,12288,0.04511377877659268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,4096,1.6090266969468858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,7168,0.01607911123169793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,4096,0.010438222024175854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,6144,0.3037351237403022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,512,0.004321777572234471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,1024,0.020624000165197585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,65536,2.562751134236654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,1536,0.0074035558435651995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,1536,0.020821332931518555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,16384,0.022640888889630634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,256,0.006701333241330252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,128,0.029498666524887085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,1536,0.07640622059504192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,768,0.010240000155236986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,10240,0.013901333014170328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,10240,0.048215111096700035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,12288,0.0395431121190389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,1024,0.003953777667548922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,12288,0.026670222481091816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,768,0.05802400244606865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,12288,0.015998222761683993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,64,0.0076142218377855085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,7168,0.5554800033569336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,12288,0.43506489859686953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,16384,0.0610391100247701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,256,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,2048,0.01331733332739936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,768,0.006296000132958095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,12288,0.21034133434295654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,3584,0.01200355589389801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,3072,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,65536,0.012053333222866058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,2560,0.015983111328548856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,51200,0.14369600349002415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,5120,0.017687110437287223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,16384,0.23517955674065483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,1536,0.02852355440457662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,12288,0.09932355748282538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,4096,0.08982221947775947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,12288,0.19807199637095133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,65536,1.4062764909532335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,512,0.009777777724795872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,1536,0.006443555570311016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,7168,0.020991111795107525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,12288,0.09867022434870402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,6144,0.049175110128190785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,5120,0.012257777982287936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,1024,0.008399110701349046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,3584,0.012619555824332766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,768,0.00433955548538102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,2048,0.03894755575391982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,4096,0.014104000396198697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,4096,0.02113688819938236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,5120,0.04695733388264974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,2048,0.012554667062229581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,8192,0.019119999474949308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,10240,0.3398551146189372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,256,0.0035111109415690103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,768,0.019073777728610568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,3584,0.00906577789121204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,7168,0.045949333243899874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,1536,0.009703111317422655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,512,0.012159999873903064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,16384,0.03588533401489258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,7168,0.6056782404581705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,12288,0.02591200007332696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,128,0.006033777776691649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,12288,0.0912346641222636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,1536,0.0131217779384719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,10240,0.03538577755292257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,8192,0.03473511007097032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,8192,0.46589157316419816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,7168,0.18370132976108125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,1536,0.004952888935804367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,2560,0.007452444069915348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,51200,0.23801422119140625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,1024,0.006014222072230444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,65536,55.247724745008675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,65536,0.11123200257619222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,128,0.014686221877733866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,3072,0.09636088874604966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,7168,0.03848177856869168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,16384,0.035191999541388616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,3584,0.006878222028414409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,7168,0.21605066458384195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,256,0.017163554827372234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,256,0.00305600009030766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,256,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,6144,0.11687910556793213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,512,0.011561777856614856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,8192,0.039790223042170204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,64,0.0047600000268883174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,51200,1.335860464307997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,65536,0.12828889158036974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,768,0.005119110975000593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,4096,0.10911466677983601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,16384,0.46672619713677305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,65536,0.08143644200430976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,7168,0.02647200061215295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,4096,0.13963733779059515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,3584,0.025769778423839148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,3072,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,1024,0.022571555442280237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,1536,0.05016888843642341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,128,0.00815466625822915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,2048,0.013919111755159167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,256,0.00425866660144594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,65536,0.7894790967305502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,64,0.17143111758761934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,3072,0.1790177822113037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,768,0.004982222285535601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,128,0.04528622163666619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,3072,0.11989511383904351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,64,0.0053137776752312975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,1536,0.1814862224790785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,10240,0.11601332823435466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,2048,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,128,0.004335999902751711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,8192,0.03169066707293192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,8192,0.010586666564146677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,3072,0.018637332651350234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,51200,0.12235377894507514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,8192,0.08540355496936375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,8192,0.45081244574652773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,5120,0.7833662033081055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,2048,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,2048,0.009165333376990424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,1024,0.011483555866612328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,1024,0.39235732290479874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,2048,0.011113777756690979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,256,0.005286222116814719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,10240,0.03740533192952474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,1024,0.018179555733998615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,4096,0.013912000589900546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,3584,0.06390933195749919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,512,0.009230222139093611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,3584,0.03348711133003235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,256,0.003683555457327101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,65536,2.3787013159857855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,5120,0.06836977932188246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,512,0.004122666600677702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,8192,0.014560000763999091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,8192,0.016210666961140104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,51200,0.08757599857118394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,51200,0.25419288211398655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,5120,0.017283555534150865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,10240,0.04081599911053976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,3072,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,8192,0.1120106644100613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,12288,1.0019671122233074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,32,0.007312888900438945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,128,0.005607111172543631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,4096,0.07709244224760267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,10240,0.06584178076850043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,2560,0.017621333400408428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,7168,0.10453333457310994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,65536,0.13711822032928467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,51200,0.358592881096734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,10240,0.30020801226298016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,2048,0.04407022065586514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,1536,0.1944417821036445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,768,0.005949333310127258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,2560,0.012245333029164208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,64,0.018990221950742934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,51200,0.33011555671691895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,64,0.0041128889554076726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,2560,0.014077333940400017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,128,0.006562666760550604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,32,0.00775733341773351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,128,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,2048,0.01841600073708428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,10240,0.013983110586802164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,8192,0.013873777455753751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,1024,0.005924444645643234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,2048,0.06990310880872938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,3072,0.05016888843642341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,256,0.03943644298447503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,51200,15.905007256401909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,3072,0.16602578428056505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,768,0.450126224093967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,768,0.015215999550289579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,5120,0.06559111012352838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,256,0.007996444072988298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,1536,0.007282666862010956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,1024,0.007093333535724216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,512,0.007855110698276097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,8192,0.03276800115903219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,64,0.01408088869518704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,128,0.015296889675988091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,10240,0.17885155147976348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,512,0.04132533404562209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,1536,0.005923555543025334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,5120,0.016904888881577384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,10240,0.021564443906148274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,8192,0.01254755589697096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,1536,0.03834222091568841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,3584,0.06826577583948772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,51200,2.973316404554579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,12288,0.10896711217032538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,2048,0.016397333807415433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,4096,0.018227555685573153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,51200,0.3126808802286784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,12288,0.021223111285103693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,10240,3.6311395433213978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,4096,1.627494176228841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,12288,0.05537511242760552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,5120,0.020688888099458482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,32,0.006566222343179915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,7168,0.02786133355564541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,8192,0.28355111016167533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,3584,0.01443733274936676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,128,0.0038088887102074097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,65536,0.8420426580641004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,7168,0.035312000248167254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,10240,0.1540106667412652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,8192,0.025584888127115037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,3072,0.009080000221729279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,10240,0.030819555123647053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,256,0.0029680000411139596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,4096,0.012474666866991254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,16384,0.02387111054526435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,256,0.004320888883537716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,256,0.0031439999325407874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,128,0.005421333428886201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,3072,0.012530666258600024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,8192,0.552528010474311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,8192,0.035559111171298556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,32,0.007965332931942409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,1536,0.0203813330994712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,3584,0.04634311132960849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,12288,0.329366233613756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,2048,0.00637155564294921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,2048,0.14518222543928358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,256,0.09513244363996719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,512,0.01514755520555708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,8192,0.12942577732933894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,8192,6.634657965766059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,12288,0.03764355513784621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,2560,0.009855111440022787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,32,0.04938755432764689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,8192,0.15275643931494817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,768,0.032948444286982216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,128,0.006492444624503453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,3584,0.015676443775494892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,1536,0.07107200225194295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,8192,0.08227288722991943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,64,0.005601777798599667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,12288,0.021719111336602107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,2560,0.020415110720528495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,1536,0.13699021604326037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,2048,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,12288,0.04189599884880913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,4096,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,64,0.0029564443975687027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,12288,0.25627467367384166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,3584,0.03303111261791653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,2560,0.008608889248636033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,10240,0.011160888605647616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,7168,0.041793776883019336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,3072,0.00675022229552269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,2048,0.006945778098371293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,3584,0.008765333228641087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,1536,0.18893067042032877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,7168,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,6144,0.019377777973810833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,6144,0.10989244116677178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,5120,0.011767110890812345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,8192,0.011734222372372946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,768,0.00423200014564726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,256,0.005564444594913059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,8192,0.06229066848754883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,5120,0.009431111315886179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,768,0.20785155561235216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,1024,0.032009777095582746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,51200,0.15963733196258545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,256,0.008512000242869059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,3072,0.03707999984423319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,2560,0.006601777755551868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,65536,0.08092711369196574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,6144,2.158618715074327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,2048,0.01260355528857973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,16384,0.055142223834991455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,1536,0.08699466784795125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,7168,0.014171555638313293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,2560,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,32,0.012268444730175866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,3584,0.6507031122843424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,12288,0.05603910817040337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,256,0.018811555372344125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,2048,0.019661333825853135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,10240,0.01607466737429301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,7168,0.0351128876209259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,65536,0.2509591049618191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,1536,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,16384,0.3793955643971761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,12288,0.2174639966752794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,2048,0.025480000509156123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,6144,0.015954666667514376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,51200,0.12339822451273601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,7168,0.0069120000633928515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,1536,0.01868355605337355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,2560,0.012337778177526263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,4096,0.007892444729804993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,8192,0.012670222255918713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,16384,0.021910222040282354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,32,0.00739822205570009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,2048,0.01036177741156684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,128,0.0041511112617121804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,2560,1.146882692972819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,64,0.003108444520168834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,32,0.013394667042626275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,6144,0.04152444336149428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,2560,0.023896000451511804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,3584,0.14772355556488037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,1536,0.006706666615274217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,1024,0.005469333380460739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,16384,0.05271111263169182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,1536,0.006525333142942852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,32,0.0027297778675953546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,3072,0.047535998953713306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,3584,0.4728079901801215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,16384,0.03736977775891622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,512,0.005923555543025334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,128,0.01956266661485036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,51200,0.1654693285624186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,51200,18.84189181857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,64,0.0605600012673272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,6144,0.0498942236105601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,512,0.015398222539159985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,8192,0.07320177555084229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,3072,0.01677688956260681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,65536,0.9898302290174695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,3072,0.018353778454992507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,6144,0.05417689018779331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,51200,0.05352888835801018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,12288,0.03103111187616984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,4096,0.058917330371008984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,4096,0.45940266715155703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,512,0.012596444123321109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,12288,0.7072533501519097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,2560,0.08873777919345432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,51200,1.8899777730305989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,1536,0.010436444646782346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,2560,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,256,0.036023110151290894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,51200,0.8822871314154731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,6144,0.2524977789984809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,1536,0.02509422269132402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,1536,0.033719112475713096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,51200,0.09726933638254802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,3584,0.014931556251313953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,512,0.005223999834722943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,1024,0.07441955804824829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,16384,0.09036088652080959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,10240,0.01865511139233907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,51200,0.25121066305372447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,12288,0.29671910074022084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,51200,0.27785245577494305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,768,0.01570133368174235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,3584,0.08236622148089938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,6144,0.012562666502263812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,7168,0.014333332578341166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,768,0.031467556953430176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,2048,0.013644444445768992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,32,0.0026844443960322272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,768,0.02569777766863505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,6144,0.011609777808189392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,6144,0.025177778469191656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,16384,0.030624889665179785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,2560,0.009843555589516958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,65536,2.3545875549316406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,32,0.00776711106300354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,256,0.004457777573002709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,64,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,7168,0.055256889926062695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,65536,0.2816942267947727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,8192,0.27900444136725533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,512,0.0033866665843460294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,2560,0.06863200002246432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,64,0.0026737778551048706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,768,0.006440889090299606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,2560,0.007127111156781514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,6144,0.14070844650268555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,256,0.02113866640461816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,256,0.027274666561020747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,256,0.028188443846172754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,7168,0.019839111301634047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,12288,0.04205422268973457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,3072,0.010243555737866296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,128,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,16384,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,8192,0.4030533366733127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,128,0.014096889230940076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,65536,7.999758402506511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,4096,0.03322133421897888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,6144,0.02731200059254964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,51200,0.23189332750108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,10240,0.27165243360731334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,2560,0.011932444241311816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,768,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,768,0.11222133371565078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,256,0.00741244438621733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,32,0.0031893334041039148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,16384,0.11090666717953152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,51200,0.9699217478434244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,8192,0.32472001181708443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,128,0.039094222916497126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,2048,0.006937777830494775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,65536,0.5076577928331163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,3072,0.01568000018596649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,4096,0.00998222248421775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,128,0.0069857777820693115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,7168,0.031025777260462444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,4096,0.011252444651391772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,6144,0.006721777634488211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,768,0.30398133065965444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,16384,0.021564443906148274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,12288,0.0517137779129876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,6144,0.1855271127488878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,65536,0.282641781700982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,10240,0.0452551113234626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,5120,0.021558221843507554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,2048,0.01556622154182858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,1024,0.005848000033034219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,512,0.4623546600341797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,2048,0.05446755554940966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,7168,3.690825568305122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,256,0.0038817777401871155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,1536,0.05079022381040785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,64,0.0031306667046414483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,16384,0.8741191228230795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,10240,0.026625777284304302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,3072,0.11822488572862412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,3072,0.007661333514584436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,512,0.005889777921968036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,12288,0.10867110888163249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,51200,0.09992088874181111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,3584,0.03875644339455499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,32,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,3584,0.03838488790724013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,2048,0.016879111528396606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,16384,0.00962311112218433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,1024,0.07865155405468412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,5120,0.018833777970737882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,32,0.008088000118732452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,2048,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,6144,0.046133332782321505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,2048,0.06921155585183038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,2048,0.7574302355448405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,51200,0.19331467151641846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,16384,0.1069271100891961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,7168,0.018571555614471436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,768,0.00423555572827657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,64,0.0028453332682450614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,2560,0.03140799866782294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,12288,0.04329599936803182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,2560,0.03391733434465196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,16384,0.3507644335428874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,12288,0.019864888654814828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,12288,0.326803551779853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,3072,0.012465777496496836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,64,0.009200000100665623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,3072,0.05403733253479004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,768,0.022115555074479844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,1024,0.005526222288608551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,256,0.0052915554907586836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,7168,0.0466577774948544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,3072,0.102528002527025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,3072,0.010915555887752108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,512,0.0047022220161226065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,512,0.004391111019584867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,768,0.006241777704821692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,6144,0.012224888636006249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,10240,0.03029688861634996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,2048,0.06064889166090223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,7168,0.021881777379247878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,6144,0.024033778243594702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,4096,0.11600266562567817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,32,0.002971555623743269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,3072,0.04264622264438205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,1536,0.03494044476085239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,51200,0.059771557648976646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,5120,0.011123555401961008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,6144,0.012612444659074148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,512,0.020225778222084045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,12288,0.024503111839294434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,768,0.005537777725193236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,2560,0.009904888768990835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,10240,0.09915822082095677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,5120,0.008284444610277811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,64,0.19599378108978271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,128,0.0053839998112784485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,16384,0.05153333478503757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,1024,0.009212444225947062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,6144,0.03849244448873732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,32,0.0036951113078329298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,64,0.5362835460238987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,1536,0.018589332699775696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,65536,0.06619022289911906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,1024,0.005187555733654234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,512,0.0033048888047536216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,2048,0.004949333353175057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,8192,0.011709333293967776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,512,0.005487999982304043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,16384,0.05451644460360209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,64,0.002558222247494592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,512,0.003913777983850903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,65536,0.1662586662504408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,8192,0.25343200895521373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,6144,0.4116640090942383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,1024,0.061576002173953585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,3584,0.018427555759747822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,3584,0.015823110938072205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,2048,0.029091556866963703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,2560,0.007366221812036302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,768,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,65536,0.11025777790281509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,7168,0.07618666357464261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,1024,0.06479644775390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,3584,0.009766221874290043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,2048,0.010593777729405297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,3072,0.048672000567118325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,3584,0.017238222890430026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,6144,0.14982577164967856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,3584,0.0178195552693473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,3072,0.4723671277364095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,65536,0.022690667046440974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,1024,0.026478222674793665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,256,0.03805689016977946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,51200,0.2436444494459364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,1536,0.035643554396099515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,128,0.010245333115259806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,65536,0.1260933346218533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,6144,0.04124355647299025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,2048,0.01053155544731352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,2560,0.012060444388124677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,768,0.016731555263201397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,12288,0.5525973108079698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,8192,2.4888701968722873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,512,0.034501334031422935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,12288,0.26586932606167263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,768,3.5796178181966147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,10240,0.04091377721892463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,4096,0.17131733894348145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,32,0.01354311075475481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,32,0.004211555752489301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,3072,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,128,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,1536,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,512,0.005971555494599872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,7168,0.012511110968059964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,51200,0.13603288597530788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,65536,0.32949333720737034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,2048,0.049357334772745766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,64,0.008408889174461365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,4096,0.01350400017367469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,6144,0.12802044550577799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,3072,0.047353777620527476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,4096,0.05228088961707222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,51200,0.06102133459515042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,51200,0.19129955768585205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,10240,0.01607022186120351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,768,0.05411822266048855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,128,0.010599999792046018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,64,0.005399999933110342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,3584,0.014495111174053617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,2048,0.006728888799746831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,5120,0.07529777950710721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,128,0.005620444400442972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,4096,0.017179555363125272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,2048,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,5120,0.7054924435085721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,16384,0.40387733777364093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,64,0.003018666679660479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,8192,0.06341866652170818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,8192,0.051721778180864125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,1024,0.025832888152864244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,6144,0.22076088852352568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,3584,0.40378578503926593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,8192,0.02590755621592204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,65536,0.5973244243197972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,12288,2.4539928436279297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,2048,0.014680888917711047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,4096,0.0759048859278361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,6144,0.040688001447253756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,16384,0.3001146581437853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,16384,0.29828977584838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,128,0.010297777752081553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,6144,0.02511288887924618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,3584,0.021368889345063105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,3072,0.11925333076053196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,128,0.00408000002304713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,2048,0.012309333516491784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,3584,0.11625511116451687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,5120,0.022562666071785822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,64,0.003080888961752256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,10240,0.03763466742303636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,128,0.021948445174429152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,256,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,64,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,65536,0.08333866463767158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,4096,0.008213332957691616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,4096,0.009008888569143083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,512,0.00518488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,1024,0.4224266741010878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,16384,0.05443200137880114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,10240,0.026734222968419392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,5120,0.014255111416180929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,8192,3.6274469163682728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,8192,0.01279022213485506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,4096,0.03011911114056905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,8192,0.13445956177181667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,256,0.005314666777849197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,128,0.002977777686383989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,51200,0.13065244091881648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,10240,0.1560266680187649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,2048,0.045372443066702954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,3584,0.022281777527597215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,1536,0.007559110720952352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,6144,0.011776000261306763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,7168,0.389756441116333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,2048,0.02160444524553087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,32,0.0029760001020299066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,1024,0.006849777781301075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,6144,0.05576088693406847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,3072,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,768,0.004967111266321606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,16384,0.006623999940024481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,8192,0.022651554809676275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,128,0.005477333234416113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,768,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,12288,0.015751111838552687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,51200,0.18908445040384927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,2560,0.14009066422780356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,768,0.009313777916961247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,512,0.004752000172932942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,8192,0.3100168969896105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,64,0.004001777619123459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,256,0.005192000004980299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,16384,0.3021706740061442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,2048,0.005967111223273807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,256,0.00655377780397733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,8192,0.04129066732194688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,6144,0.035615109735065036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,51200,0.8770595656500922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,3584,0.012216889195972018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,6144,0.026509332987997267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,51200,0.15573244624667698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,6144,1.3797867033216689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,6144,0.011165333290894827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,10240,0.03373600045839945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,8192,0.013030221892727746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,7168,0.014196443888876172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,256,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,128,0.004967111266321606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,5120,0.034830222527186074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,2560,0.0875226656595866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,12288,0.6668746736314562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,12288,2.369150161743164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,10240,0.22444534301757812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,256,0.004035555654101902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,5120,0.011634666886594562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,65536,1.2407511605156791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,512,0.3460053337944879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,12288,0.5191849072774252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,65536,0.14235466056399876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,768,0.004648888690604104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,5120,0.015644444359673392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,1024,0.005557333429654439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,7168,0.014475555883513557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,1024,0.008794666992293464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,256,0.009550222092204625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,2048,0.010772444307804108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,4096,0.026130666335423786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,2048,0.010438222024175854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,768,0.010372444159454769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,32,0.0034764446318149567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,16384,0.07008888986375597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,1536,0.010650667051474253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,8192,0.026164443956481084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,1536,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,12288,0.05840266413158841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,1024,0.010264000131024255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,7168,0.026915555198987324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,256,0.003253333270549774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,12288,0.4711902406480577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,128,0.005009777843952179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,3584,0.2122053305308024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,6144,0.013792888985739814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,32,0.16289956039852566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,1024,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,6144,0.11027999718983968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,128,0.03974666529231601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,128,0.007724444071451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,7168,0.015213333898120456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,3584,0.019663110375404358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,8192,0.0841733349694146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,5120,0.032447109619776406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,3072,0.08373510837554932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,10240,0.01370666672786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,1536,0.018601778480741713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,3072,0.05054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,3072,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,256,0.0033404444240861465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,32,0.0028782222006056043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,10240,0.27493778864542645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,1536,0.011289777855078379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,51200,0.3078062269422743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,2560,0.075527999136183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,6144,0.01974933346112569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,5120,0.11828445063696967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,2560,0.006239111224810283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,12288,0.011801777614487542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,16384,0.008370666868156856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,1536,0.004401777767472797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,4096,0.14036533567640516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,4096,0.06932710938983493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,128,0.0026568888376156488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,2048,0.00793511089351442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,3584,0.007837333612971837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,8192,0.020222221811612446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,12288,0.05205333232879639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,3584,0.024933333198229473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,256,0.00489777781897121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,3072,0.04830044507980347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,6144,2.7666204240587025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,3584,0.09391378031836616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,5120,0.016396444704797532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,2560,0.10128621922598945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,768,0.004765333400832282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,3072,0.006238222122192383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,12288,0.0701342225074768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,10240,0.5001031027899848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,3584,3.0662453969319663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,12288,1.315233760409885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,5120,0.015652444627549913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,3072,0.01145066652033064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,3072,0.16996888319651285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,2560,0.009678222239017487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,512,0.005621333503060871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,1536,0.029079112741682265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,256,0.006476444502671559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,512,0.019900444481107924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,768,0.11722577942742242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,8192,0.025255110528733995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,16384,0.059261335266960986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,768,0.004305777864323722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,32,0.003630222131808599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,7168,0.023465777436892193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,6144,0.18034222390916613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,6144,0.02148977749877506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,768,0.004390222330888112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,7168,0.011141333315107556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,1024,0.021632888250880774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,10240,0.056576000319586865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,2560,0.021021333005693223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,2048,0.005264889034960005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,12288,0.01754399968518151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,3072,0.051983998881445996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,5120,0.06404355499479505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,10240,0.08049511247211032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,1536,0.048087110122044883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,4096,0.010515555739402771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,5120,0.27793600824144155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,6144,2.9505591922336154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,12288,0.2683111031850179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,51200,0.13043288389841715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,64,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,128,0.005903999838564131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,768,0.005700444595681296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,128,0.005633777628342311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,3072,0.010316444767846001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,1536,0.08371644549899632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,10240,0.03259288933542039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,512,0.011320888996124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,51200,0.05826489130655924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,12288,0.0069226668112807805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,10240,0.039678222603268094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,65536,2.955844455295139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,512,0.0074551113777690465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,1024,0.012057777908113269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,8192,0.009840889109505547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,128,0.006746666712893381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,1536,0.9814142651028104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,2048,0.0998071100976732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,3072,0.018995554910765756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,1024,0.0058551111982928375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,10240,0.22141599655151367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,65536,1.6218836042616103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,16384,0.8033226860894097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,3072,0.00927022182279163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,6144,0.01921688848071628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,16384,0.7966791258917915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,16384,0.08368800083796184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,16384,0.14452889230516222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,12288,0.02335111134582096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,768,0.05197688937187195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,6144,0.41501421398586696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,2560,0.01167022188504537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,32,0.005398222141795688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,4096,0.01847466660870446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,768,0.014840000205569796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,12288,0.9912213219536675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,64,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,6144,0.022188444932301838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,6144,0.01644622286160787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,65536,0.07558400101131864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,512,0.08020266559388903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,1024,3.190630171034071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,12288,0.0074311114019817775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,1024,0.00600888869828648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,256,0.010176888770527309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,7168,0.1251351171069675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,3072,0.012878222597969903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,2560,0.0076044441925154785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,2048,0.0075146663520071245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,10240,9.880957709418402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,10240,0.27329956160651314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,3072,0.014717333846622042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,5120,0.012858666479587555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,4096,0.013781333135233985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,2560,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,16384,5.856494903564453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,65536,0.1406408945719401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,12288,0.3057653374142117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,3072,0.020624000165197585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,768,0.010654222634103564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,3584,0.03154133425818549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,1024,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,128,0.0029404444826973807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,3584,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,12288,0.1424364381366306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,5120,0.4545760154724121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,6144,0.022846221923828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,3072,0.04389244318008423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,12288,0.33773332171969944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,128,0.003014222201373842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,2560,0.11501777172088623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,1536,0.005761777775155173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,2048,0.006562666760550604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,1024,0.04549511273701986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,10240,0.04276622335116068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,3584,0.00851111114025116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,51200,0.7405626508924695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,2560,0.25671201282077366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,12288,0.18062133259243437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,8192,0.15502222379048666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,5120,0.03869333200984531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,51200,0.05927733580271403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,64,0.002495999965402815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,32,0.0033822223130199644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,1024,0.04666755596796671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,16384,0.28362398677402073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,64,0.0026293332791990707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,8192,0.22460532188415527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,1024,0.019136000010702345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,512,0.004934222333961063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,64,0.03544266687499152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,64,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,6144,0.021869333253966436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,10240,0.028095109595192805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,6144,0.015856888559129503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,8192,0.015942222542232938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,768,0.003670222229427762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,1536,0.005381333331267039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,12288,0.21068533261617026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,7168,0.03145600027508206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,5120,0.01638488968213399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,6144,0.0693919989797804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,2048,0.0684702197710673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,4096,1.6407848993937175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,3584,0.07753155628840129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,512,0.011372444530328115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,4096,0.03000266684426202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,256,0.0052177777720822235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,4096,0.009597332941161262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,128,0.004986666556861666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,51200,0.3149831030103895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,128,0.028569777806599934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,256,0.0032844444115956626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,51200,0.8900213241577148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,128,0.016920000314712524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,6144,0.021732444564501446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,1536,0.4046444363064236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,12288,0.6024879879421657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,51200,0.05091199941105313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,1536,0.011387555963463254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,65536,0.03941422369745042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,7168,0.014800000521871777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,7168,0.040158222119013466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,16384,0.3493395646413167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,512,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,5120,0.7491697735256619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,768,0.04959466722276476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,1536,0.009166222479608323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,4096,0.020284444093704224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,32,0.013223999904261695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,7168,0.1483466625213623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,16384,0.6238320138719347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,8192,0.14239733748965794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,3584,0.300930658976237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,5120,0.011221333510345884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,4096,0.016233777006467182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,51200,0.045751998821894325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,7168,0.49655108981662327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,51200,0.06781066788567437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,16384,0.06578310992982653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,16384,0.1883253388934665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,3584,0.05784533421198527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,256,0.00628355559375551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,64,0.06693955262502034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,32,0.002761777697337998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,6144,0.008738666772842407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,3072,0.025384000606007043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,12288,0.3510275681813558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,7168,0.012039999994966718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,32,0.005375111268626318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,64,0.004959110998445087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,256,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,128,0.009148444566461775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,51200,0.7252186669243706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,6144,0.033730665842692055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,1536,0.007453333172533248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,16384,0.0745484431584676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,1536,0.004601777841647466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,768,0.005678222411208683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,2560,0.09603822231292725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,65536,0.08072533210118611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,16384,0.08055466413497925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,65536,0.14863555961185032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,2048,0.006204444501135085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,10240,0.08499199814266628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,51200,1.327745755513509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,8192,0.020304888486862183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,2560,0.06005066633224487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,10240,0.4619715478685167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,10240,0.039334221018685236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,3584,0.009720889230569204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,8192,0.4712622430589464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,768,0.006487111250559489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,12288,0.04466933343145582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,32,0.0032799999333090256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,128,0.018876444962289598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,256,0.014057776994175382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,2048,0.01569422251648373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,51200,0.7077848646375867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,2560,0.02086666723092397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,12288,0.021690666675567627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,10240,0.18984533680809867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,1536,0.005274666680230035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,768,0.04268444577852885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,1024,0.012469333079126147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,3072,0.008959111240175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,65536,0.35142132971021867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,1536,0.018177777528762817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,7168,0.2450693448384603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,3584,0.01779822177357144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,256,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,8192,0.009954666925801171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,768,0.004572444491916233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,16384,0.7684826850891113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,5120,0.02535555594497257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,2560,0.07634844382603963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,4096,0.07472266753514607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,2560,0.0669599970181783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,6144,0.010127999716334874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,512,0.004089777668317159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,256,0.00480000012450748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,768,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,6144,0.16745332876841226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,32,0.2885804441240099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,256,0.009741333623727163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,3584,0.034443553951051496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,4096,0.11079111364152695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,12288,0.7804462114969889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,10240,0.019679110911157396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,2048,0.008540444076061249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,32,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,256,0.007903999752468532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,128,0.023736889163653057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,3584,0.008188444707128737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,8192,0.2256151040395101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,6144,0.22399822870890299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,16384,0.155330671204461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,4096,0.01128355579243766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,12288,0.020080000162124634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,64,0.005524444497293896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,5120,0.029772443903817072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,2048,0.17957777447170684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,1536,0.013791110780504016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,32,0.012416889270146688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,51200,0.1906728876961602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,768,0.011863110793961419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,7168,0.18537155787150064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,1536,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,1024,0.009848888549539778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,51200,0.054582221640480884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,3072,0.198416895336575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,6144,3.9939617580837674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,51200,0.11014399925867717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,12288,0.10604888863033718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,2560,1.2956755956013997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,7168,0.014808000789748298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,32,0.006769777586062749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,2560,0.013334222137928009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,16384,0.05585955248938667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,256,0.005425777700212266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,2560,0.8186035686069064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,256,0.004839999808205498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,1536,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,6144,0.014225777652528552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,51200,1.0089715321858723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,7168,0.02122666603989071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,768,0.09263644615809123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,6144,0.02492888934082455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,512,0.005271111097600725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,2048,0.04203644394874573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,16384,0.030976888206270006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,3072,0.015250666273964776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,128,0.03931466738382975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,256,0.012494222157531314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,5120,0.06290666924582587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,12288,0.060737775431738965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,256,0.0035333335399627686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,64,0.007166222565703922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,32,0.005869333528810077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,4096,0.028930667373869155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,10240,0.017235555582576327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,512,0.2479164335462782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,6144,0.1517039934794108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,5120,0.07093244128757052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,5120,0.030943999687830608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,1536,0.029055999384986028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,768,0.020011555817392137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,3072,0.06609244479073419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,6144,0.011191999746693505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,10240,0.02233955595228407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,1536,0.03312888741493225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,32,0.17630310853322348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,65536,0.012545777691735161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,5120,0.008193777667151557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,256,0.00563822231358952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,6144,0.016392888294325936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,10240,0.3897582160101996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,768,0.00433955548538102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,65536,0.19626311461130777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,3072,0.037554668055640325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,256,0.004482666651407878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,3584,0.16812977525922987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,6144,0.03772088885307312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,64,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,2048,0.029333333174387615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,6144,0.031382222970326744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,16384,0.02530844509601593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,3072,0.2764151096343994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,8192,0.03363022208213806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,3584,0.05808266666200426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,10240,0.025430222352345783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,7168,0.1090631087621053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,512,0.04990488953060574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,7168,0.012288889123333825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,1024,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,128,0.04340622160169813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,51200,0.06784800026151869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,2560,0.010299555957317352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,8192,0.03976800044377645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,2048,0.035691556003358625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,10240,2.0233164893256292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,512,0.007118222614129384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,64,0.13800888591342503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,4096,0.010948444406191507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,3584,0.009062222308582729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,6144,0.023919999599456787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,4096,0.01921333372592926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,12288,0.09685689210891724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,65536,0.12753421730465359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,3584,0.021368889345063105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,2560,0.022673777408070032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,768,0.0037271111375755737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,128,0.009297777381208207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,768,0.020095111595259774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,512,0.00545155546731419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,2048,0.386424011654324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,8192,0.022237333986494277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,7168,0.01887022289964888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,512,0.027066666218969557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,2048,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,2048,0.04084622197681003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,16384,0.025062221619817946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,768,0.003581333491537306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,12288,0.045534223318099976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,51200,0.8824311362372504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,2048,0.13622400495741102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,5120,0.16909778118133545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,12288,0.03676088982158237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,2048,0.02219111058447096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,6144,0.0718088878525628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,512,0.009522666533788046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,51200,0.3508675628238254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,12288,0.744067562950982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,2560,0.01310400002532535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,64,0.029291556941138372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,1024,0.12166933218638103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,10240,0.19747377766503227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,16384,0.012703110774358114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,51200,0.012862222062216865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,6144,0.09746578004625107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,1536,0.0063271112740039825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,3584,0.029511110650168523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,51200,0.09159111314349705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,768,0.003834666477309333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,5120,0.009372444616423713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,128,0.0027155555370781156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,16384,0.524336020151774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,1536,0.013125333521101209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,128,0.005502222312821283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,512,0.005130666825506422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,16384,0.43019734488593203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,1024,0.04773599902788798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,12288,0.04522222280502319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,1536,0.09939644734064738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,6144,0.03453155689769321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,3584,0.040975110398398504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,4096,0.230841769112481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,8192,0.04211911227968004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,4096,0.045371555619769625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,3072,0.011004444625642566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,3584,0.02534488836924235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,512,0.009546666509575313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,51200,0.13655022780100504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,128,0.011063111325105032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,2560,0.8422462145487467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,512,0.01647822227742937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,7168,0.03831911087036133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,5120,0.01421511173248291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,7168,0.01947733263174693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,5120,0.03635466761059231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,4096,0.27883556154039174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,512,0.008904000123341879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,7168,0.017675555414623685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,256,0.05983022186491224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,3072,0.021934222843911912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,3072,0.3267928759256999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,128,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,51200,0.06941866874694824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,16384,0.025607999828126695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,1024,0.011197333534558615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,3072,0.009128000173303816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,6144,0.02386577758524153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,6144,0.1980311075846354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,8192,0.006525333142942852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,4096,0.10273422135247125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,10240,0.36734578344557023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,128,0.027983112467659846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,32,0.0027306667632526825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,7168,0.6328738000657823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,16384,0.0262773334980011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,10240,0.35988712310791016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,1536,0.3400604459974501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,51200,2.1400985717773438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,5120,0.0889822244644165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,8192,0.22217955854203966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,51200,0.1657137738333808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,2560,0.00739733295308219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,3584,0.007065777977307637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,3072,0.009504000345865885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,64,0.006821333534187741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,3584,0.08736622333526611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,8192,0.02446311049991184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,6144,0.04995377858479818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,64,0.0029768887907266617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,1536,0.005449777675999536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,16384,0.07509422302246094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,2560,0.13312799400753444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,128,0.005199111170238919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,3072,0.006801777829726537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,768,0.005768888940413793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,16384,0.019360888335439894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,256,0.011262222296661802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,5120,0.018361777067184448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,32,0.0033200000309281876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,12288,0.1382675568262736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,512,0.02271555529700385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,4096,3.087888929578993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,1536,0.054899556769265064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,3584,0.011556444068749746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,16384,0.3565662172105577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,12288,0.06297422117657132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,768,0.0151573336786694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,6144,0.017867555220921833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,1024,0.010636444720957013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,32,0.005492444253630108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,10240,0.12834489345550537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,1536,0.0047564444442590075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,8192,0.03010666701528761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,7168,0.017839999662505258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,3072,0.013765333427323235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,7168,6.121426476372613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,2048,0.009648888475365108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,1536,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,51200,2.5206756591796875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,32,0.0057839999596277876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,1024,0.026571555270089045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,8192,0.05931466817855835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,1536,0.10880088806152344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,65536,0.12893066141340467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,16384,0.11796177758110894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,3584,0.012646222280131446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,12288,0.01570399933391147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,256,0.004553777890072929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,512,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,512,0.007858666280905405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,5120,0.43575467003716367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,2048,0.007975999679830339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,16384,0.04387111133999295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,51200,0.15731910864512125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,256,0.022638221581776936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,65536,0.423001766204834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,16384,0.024842666255103216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,1536,0.011926222178671094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,7168,0.011159111228254108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,256,0.004722666823201709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,512,0.0069413334131240845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,768,0.010934222075674268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,6144,0.02790133489502801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,5120,0.10376533534791733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,512,0.014398222168286642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,128,0.006833777659469181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,3072,0.02295111119747162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,10240,0.01835911141501533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,6144,0.33318932851155597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,64,0.015811555915408664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,1024,0.00600888869828648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,3584,0.009365333451165093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,4096,0.30424178971184623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,256,0.010888888604111142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,51200,0.9279084735446506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,10240,0.2620791064368354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,128,0.009973333113723332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,64,0.0026026666164398193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,3072,0.02286311156219906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,8192,0.181658665339152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,32,0.005630222045713001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,4096,0.045078221294615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,16384,0.08758844269646539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,6144,0.01774222155412038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,128,0.03138400117556254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,5120,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,8192,0.059782220257653125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,8192,0.16926222377353242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,4096,0.035113778379228376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,5120,0.012888000243239932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,3584,0.008316444853941599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,768,0.016775111357371014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,12288,0.06514044602711995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,1536,0.009053332938088311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,32,0.0026666666898462507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,768,0.01696622206105126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,64,0.009829333258999718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,3584,0.020484444167878892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,64,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,512,0.005311111195219888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,2048,0.011095110740926532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,1024,0.008937777744399177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,65536,0.10221066739824082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,768,0.004942222187916438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,512,0.013289777768982781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,1024,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,2560,0.007840000092983246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,8192,0.02642755541536543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,512,0.30537867546081543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,32,0.014663999279340109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,6144,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,2560,0.007828444242477417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,16384,0.03741688860787286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,1024,0.06750577688217163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,65536,1.095191107855903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,2560,0.009761778016885122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,12288,0.020950223008791607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,8192,0.027821332216262817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,5120,0.011023110813564725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,512,0.01772533357143402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,51200,0.11064177751541138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,5120,0.02139288849300808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,4096,0.006106666806671355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,512,0.009401777552233802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,3072,0.017088888419999015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,5120,0.06394044558207194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,2560,0.12588711579640707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,256,0.05178933342297872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,768,0.009927111367384592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,6144,0.1626746654510498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,5120,0.08233510785632663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,8192,0.08489244514041477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,3072,0.18088089095221627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,3584,0.036093334356943764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,8192,0.006616888774765863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,16384,0.022783999641736347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,32,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,5120,0.014709333578745524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,7168,0.01481599940194024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,64,0.003679111186001036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,12288,20.26785617404514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,6144,0.2608613438076443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,12288,0.07531288597318861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,3584,0.02150844368669722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,65536,0.9588070975409614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,512,0.0063591111037466265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,65536,1.1874738269382052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,32,0.002540444541308615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,51200,0.3230604330698649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,5120,0.2791626718309191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,64,0.008243555823961893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,5120,0.5394017961290147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,5120,0.009958222508430481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,1536,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,3072,0.007628444168302748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,6144,0.1476480033662584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,3584,0.8905048370361328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,10240,0.4104817708333333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,2560,0.038176000118255615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,51200,0.1061813301510281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,16384,0.4215911229451497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,16384,0.1293048858642578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,7168,0.05196533269352383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,256,0.01647822227742937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,2048,0.00961866643693712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,32,0.011125333607196808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,3072,0.03522133496072557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,256,0.018173333671357896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,10240,0.30758844481574166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,51200,0.05469422207938301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,4096,0.009672889278994666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,64,0.0046986668474144405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,1536,0.00812711069981257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,5120,0.013349333571063148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,256,0.0042151109211974675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,768,0.005993777679072485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,65536,0.27704087893168133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,1024,0.0058559998869895935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,6144,0.008847999903890822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,64,0.005552000055710475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,768,0.006032889087994893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,1024,0.00963644435008367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,512,0.00564533347884814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,768,0.005717333406209946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,7168,0.025567111041810777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,2048,0.10038844744364421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,3584,0.012315555579132505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,128,0.005029333134492238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,2560,0.09059466918309529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,64,0.003097777772280905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,5120,0.046767999728520714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,65536,0.29369955592685276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,51200,0.662608888414171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,64,0.015505777464972602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,256,0.0085235552655326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,7168,0.031750222047170006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,6144,0.01663466625743442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,64,0.03185333477126227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,128,0.033935109774271645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,65536,0.2147982252968682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,768,0.007715555528799693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,6144,0.01017599966790941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,12288,0.030991110536787245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,10240,0.35979554388258195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,64,0.006632888896597757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,7168,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,1024,0.031050665511025324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,16384,10.766560024685331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,7168,0.029995554023318823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,2048,0.07597777578565809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,10240,0.423018667432997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,32,1.5679812961154516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,5120,0.16883199744754365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,64,0.01034311122364468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,2048,0.01346666696998808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,32,0.0032595555401510666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,2560,0.04462488823466831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,3584,0.08112266990873548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,3072,0.06140622165468004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,65536,0.8601857821146647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,8192,0.03716977768474155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,65536,0.08279199732674493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,12288,0.0064328888224230866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,3072,0.019444444113307528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,6144,0.018651554981867473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,7168,0.03147466646300422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,2560,0.03888444436921014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,512,0.005499555418888728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,65536,0.3863875601026747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,4096,0.09073688586552937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,256,0.1086133321126302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,4096,0.15189066198137072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,16384,0.01796000036928389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,32,0.09310222334331936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,6144,0.14291377862294516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,512,0.0053226666318045715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,2048,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,512,0.005383111122581694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,32,0.002965333354141977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,65536,0.01514755520555708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,2048,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,4096,0.8324782053629557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,256,0.007698666718271043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,2048,0.013288000391589271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,5120,0.038980444272359215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,10240,0.009376889301670922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,7168,1.116369777255588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,12288,1.4966248406304254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,1536,0.02529688841766781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,10240,0.15260977215237087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,51200,0.14498666922251383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,8192,0.027099554737408955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,7168,0.014947555131382413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,128,0.00407911092042923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,256,0.003710222327046924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,51200,0.06373600165049235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,65536,5.416438208685981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,768,0.06983644432491727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,256,0.006984000404675801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,32,0.0034142221427626083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,10240,0.01460800071557363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,65536,0.3599395487043593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,1536,0.10241422388288711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,8192,0.040427555640538536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,3072,0.13392089472876653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,6144,0.04475733306672838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,32,0.0211191111140781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,1024,0.35907644695705837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,64,0.010832889212502373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,12288,0.023192889160580103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,65536,3.305904812282986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,512,0.006063111126422882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,512,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,65536,0.2072835498385959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,3072,0.020374221934212577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,5120,0.014007111390431723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,5120,0.009383999639087254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,10240,0.5637333128187392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,4096,0.007803555164072249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,6144,0.019997333486874897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,7168,0.21203822559780547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,3072,0.07970844374762641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,4096,0.02125688890616099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,7168,0.07426044676038954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,3584,0.011807111402352652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,3072,0.007409777906205919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,12288,0.046775112549463906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,4096,0.1545004447301229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,3584,0.7531981998019748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,1024,0.007355555891990662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,51200,0.009396444592210982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,5120,0.01108266661564509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,64,0.009026666482289633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,65536,0.21721243858337402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,3584,0.01724177764521705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,768,0.0047937776479456155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,3584,0.11560444037119548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,1536,0.006426666759782367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,8192,0.06568355692757501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,6144,0.05380444394217598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,6144,0.06338933441374037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,768,0.009672000176376766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,3072,0.06132533152898153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,512,0.005447111195988125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,1536,0.03410311208830939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,128,0.010456889039940305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,12288,0.061554663711123996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,5120,0.01239911135700014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,12288,0.026009778181711834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,3584,0.014257777068350049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,51200,0.6971431308322482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,1024,0.06483733654022217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,1024,0.010808000134097205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,65536,0.5242186652289497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,51200,0.02016799979739719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,768,0.011143111520343356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,1024,0.04505066739188301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,32,0.005449777675999536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,64,0.0057911111248864075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,64,0.007698666718271043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,12288,0.033238222201665245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,4096,0.017550221747822232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,1536,0.011651555697123209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,768,0.004957333207130432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,512,0.005372444374693765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,5120,0.20431378152635363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,6144,0.018793778287039865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,1536,0.10936178101433648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,256,0.002977777686383989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,2560,0.01052177780204349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,12288,0.7101270887586805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,4096,0.016739555531077914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,16384,0.034272889296213783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,1536,0.020822222034136455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,3584,0.24725511338975695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,3072,0.0063680000603199005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,1536,0.015263999501864115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,256,0.0041893331540955436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,3584,0.23781689008076987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,2048,0.0894506639904446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,256,0.005041777673694822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,128,0.008471999731328752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,16384,0.06769955820507474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,16384,0.021370665894614324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,51200,0.6020684242248535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,1536,0.005000888887378904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,2560,0.015462223026487561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,64,0.004973333328962326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,512,0.0034328889515664843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,1536,0.013796444568369122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,8192,0.10628177722295125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,65536,0.7846453454759387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,1536,0.013334222137928009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,512,0.016178665889634024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,7168,0.0487040016386244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,512,0.02496888902452257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,256,0.004516444272465176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,8192,0.4064497682783339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,1024,0.10256089104546441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,2560,0.030100444952646892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,65536,18.978663126627605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,12288,0.13901600572797987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,256,0.0032791110376516977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,4096,0.11005866527557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,32,0.005167110926575131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,10240,0.011413333316644033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,256,0.00443733317984475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,3072,0.02313599983851115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,5120,0.022111111217074927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,65536,4.172830369737413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,65536,0.09395111269421047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,768,0.01369066701995002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,1536,0.011828444070286222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,7168,0.23316889339023164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,512,0.12858933872646755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,64,0.016398222910033334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,3072,0.02846755584081014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,2048,0.019153777096006606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,32,0.0028417776856157514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,1024,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,128,0.0052533331844541765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,2560,0.0997048881318834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,32,0.006324444380071428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,128,0.003116444374124209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,65536,0.0702915522787306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,1024,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,512,0.00327822214199437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,16384,0.05355822377734714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,1536,0.09621155261993408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,10240,0.04420533445146349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,51200,0.20485956139034697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,2560,0.012435555458068848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,2560,0.01605955594115787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,256,0.0037884443170494507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,3072,0.007244444555706448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,8192,0.6146622233920628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,16384,0.022056889202859666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,65536,0.031165334913465712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,2560,0.013278221918476952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,768,0.028718223174413044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,51200,0.2410879929860433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,5120,0.006822222222884496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,2048,0.007968000239796108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,6144,0.06311200062433879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,256,0.02236977716286977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,8192,0.023802666200531855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,2560,0.01572088897228241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,256,0.01531733406914605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,16384,0.2734977669186062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,1024,0.006543111056089401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,51200,0.18801421589321562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,1536,0.014120888378885059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,4096,0.24292267693413627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,1024,0.02572533322705163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,7168,0.1878115601009793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,3584,0.0251875552866194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,128,0.0053280000057485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,256,0.006484444356626934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,3584,0.09669244289398193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,51200,0.9268115361531576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,1536,0.041843556695514254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,3072,0.03658311234580146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,3584,0.15135110749138728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,7168,0.0635146697362264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,3584,0.009645333720578088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,512,0.004132444245947732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,3584,0.021573333276642695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,10240,1.3803591198391383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,3584,0.01073333372672399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,7168,0.015909334023793537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,12288,0.10137066576215957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,10240,0.02695200012789832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,5120,0.0069351109365622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,768,0.02846577763557434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,4096,0.010682666467295753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,65536,0.08122399780485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,10240,0.015077332655588785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,128,0.0047404443224271136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,1024,0.01328000012371275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,51200,6.217397477891709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,64,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,7168,0.014307555225160388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,3584,0.22098310788472494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,64,0.007298666569921706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,12288,0.019556444552209642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,768,0.00655644428398874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,51200,0.4569750891791449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,2560,0.015981333123313058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,32,0.012857777376969656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,3584,0.10705333285861546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,65536,0.15021422174241808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,3584,0.2570879989200168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,65536,17.24918450249566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,65536,0.5992586877610948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,2048,0.010264889233642155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,7168,0.02644177774588267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,5120,0.006911110960774952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,16384,0.5977048873901367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,512,0.10389510790506999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,65536,0.4070453378889296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,3072,0.01601777805222405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,1024,0.010581333604123859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,1536,0.006911110960774952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,3584,1.0746338102552626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,768,0.005464000006516774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,65536,0.10033777687284683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,2048,0.01630311045381758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,3072,0.008542222281297049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,5120,0.18863733609517416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,6144,0.0116462219092581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,768,0.005913777897755305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,6144,0.017331555485725403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,256,0.006631999793979857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,1536,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,256,0.019350222415394254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,32,0.0028239999794297745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,64,0.0026666666898462507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,6144,0.1275955571068658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,1536,0.013935110635227628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,64,0.0032524443748924467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,51200,1.8128835890028212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,51200,0.15979822476704916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,1536,0.17796711126963297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,7168,0.016519111063745286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,12288,0.00831200016869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,1024,0.01296533313062456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,12288,0.3843759960598416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,5120,0.38998132281833225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,3584,0.009825777676370408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,128,0.0033048888047536216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,1536,0.0760986672507392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,1024,0.05444000164667765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,64,0.010807111031479306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,3072,0.009953777823183272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,768,0.0041644444896115195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,32,0.004603555632962121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,5120,0.015142222245534262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,2048,0.007330666813585494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,10240,0.03546488947338528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,1024,0.006122666514582104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,12288,0.08560622400707668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,1536,0.0046968890560997855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,128,0.05177422364552816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,64,0.0064604443808396654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,1536,0.006271111054552927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,128,0.0051857779423395796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,10240,0.017838221457269456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,1024,0.07446044683456421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,3072,0.047296000851525195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,128,0.022524444593323603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,5120,0.0251511103577084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,3584,2.1316755082872176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,128,0.008626666333940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,10240,0.04315733247333103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,3584,0.030219554901123047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,1024,0.004687110996908612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,10240,0.03726044297218323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,5120,0.013125333521101209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,7168,0.25010132789611816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,128,0.005699555493063397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,256,0.0034711110509104202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,768,0.004431999805900785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,7168,0.07143733236524794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,4096,0.13795910941229925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,5120,0.09691733121871948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,51200,1.1354800330268013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,16384,0.032786667346954346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,51200,0.11100000143051147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,12288,1.0655493206448026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,3072,0.010110222631030613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,6144,0.1666240029864841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,7168,0.8013182216220431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,65536,0.7183439996507434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,128,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,10240,1.0480498207939994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,768,0.2720959981282552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,128,0.004581333448489507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,16384,0.41719910833570695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,5120,0.06023110946019491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,768,0.004092444562249714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,16384,0.037259555525249906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,256,0.2153466675016615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,16384,0.06890310843785603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,256,0.00674222202764617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,3584,0.02164177762137519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,2048,0.06391644477844238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,2560,0.039956443839603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,1024,0.06910222106509738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,256,0.01865511139233907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,1024,0.03453155689769321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,2048,0.4345324569278293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,4096,0.02458488941192627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,6144,0.017699556218253244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,10240,0.013575110998418597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,16384,0.18218844466739229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,65536,1.1108079486423068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,51200,0.12368088298373753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,10240,0.01696888936890496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,5120,0.012015110916561551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,8192,0.013079110946920184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,1024,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,12288,0.02513422237502204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,64,0.010370666782061258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,2560,0.030846221579445735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,16384,0.09033688571718003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,16384,0.2404844495985243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,51200,0.08230044444402058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,1536,0.09099466933144464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,10240,0.3276008764902751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,512,0.007557333343558841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,1536,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,1024,0.008986666798591614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,1024,0.015984000431166757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,7168,0.011798222031858234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,512,0.023457777169015672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,1024,0.010483555495738983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,2560,0.058152887556287974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,2560,0.03387555480003357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,8192,0.03906044363975525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,3072,0.015604444675975375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,7168,0.03225866622394986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,1024,0.005548444473081165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,4096,0.010653333531485664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,1024,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,2560,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,32,0.011906666888131036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,768,0.0037324445115195382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,3072,0.007818666597207388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,12288,0.993093384636773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,4096,0.019899555378490023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,5120,0.024286222126748826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,65536,0.2087937725914849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,1536,0.02942133280966017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,4096,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,1024,0.05249955587916904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,128,0.008773333496517604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,3584,0.006899555524190267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,2560,0.08113688892788358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,32,0.00295733329322603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,128,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,4096,0.007948444121413762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,3584,0.011567999919255575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,51200,0.21800266371832955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,64,0.026496888862715826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,5120,0.2807493209838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,5120,0.021240888370407954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,7168,0.00868444475862715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,3072,0.011323555476135679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,32,0.07680355840259127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,64,0.006089777582221561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,10240,0.3567235469818115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,10240,0.028419554233551025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,51200,0.2697804503970676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,512,0.004648000001907349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,7168,0.21385333273145887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,4096,0.01126577787929111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,512,0.005921777751710679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,12288,1.365687158372667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,768,0.011643555429246692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,1536,0.005356444252861871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,6144,0.05076266659630669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,1536,0.029491557015313044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,12288,0.22427466180589464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,65536,13.768226623535156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,1536,0.012533333566453723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,5120,0.4585386912027995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,768,0.005602666487296422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,1536,0.008732444710201686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,7168,0.0904959970050388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,7168,0.011624000138706632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,64,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,2048,0.3724551200866699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,7168,0.0064897777305708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,12288,0.6380400127834743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,1024,0.02241866621706221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,4096,0.09844355450736152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,768,0.0038897775941424896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,6144,0.00685422205262714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,16384,1.5261991288926866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,64,0.0033857776886887024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,7168,0.27346934212578666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,512,0.009787555370065901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,4096,0.03197688857714335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,1024,0.03746933407253689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,16384,0.06832177771462335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,1024,0.02057066725360023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,256,0.0031386667655573953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,6144,0.04989333285225762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,128,0.38133777512444395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,32,0.03968977928161621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,3072,0.08146578073501587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,51200,0.03166666626930237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,512,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,7168,0.54239288965861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,7168,0.3263804382748074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,128,0.02755555510520935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,6144,0.02536533276240031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,5120,0.09080977572335137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,8192,0.01369066701995002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,512,0.02103288968404134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,64,0.005843555761708154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,3584,0.009148444566461775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,5120,2.4055858188205295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,8192,0.18272444936964247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,6144,0.03279111120435927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,128,0.02176800039079454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,10240,0.0380657778845893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,768,0.006484444356626934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,2048,0.05721155802408854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,16384,0.49288357628716367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,6144,0.03708533445994059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,768,0.03380533390574985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,12288,0.14205866389804417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,1536,0.008615111311276754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,16384,0.08595466613769531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,65536,0.25032711029052734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,256,0.003400000019205941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,65536,0.36972800890604657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,65536,2.1230852339002824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,3584,0.024596444434589807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,2048,0.0075137780772315125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,768,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,1536,0.4490559895833333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,1536,0.0075333333677715724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,512,0.004132444245947732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,128,0.003559999995761447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,10240,0.047777778572506376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,8192,0.1162453360027737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,12288,0.06838399834103055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,1536,0.02805066770977444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,2048,0.060808002948760986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,4096,0.011337777806652917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,16384,0.1824960046344333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,6144,0.010687111152542962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,12288,0.025032000409232244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,3072,1.1302826139662001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,32,0.0031022222505675424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,51200,0.07786577939987183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,256,0.005575110928879843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,2048,0.007857778006129796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,12288,0.05870311127768623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,10240,0.07947822411855061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,3072,0.01595288846227858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,3072,0.0136817776494556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,3072,0.007058666812049017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,512,0.0037884443170494507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,1024,0.0059004442559348206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,32,0.034132444196277194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,4096,0.030981332063674927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,7168,0.10825688971413507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,32,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,5120,0.8664337793986002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,512,0.013581333061059317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,1024,0.0052471111218134565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,128,0.009138666921191746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,512,0.004842666702138053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,3584,0.013502221968438892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,8192,0.17477956083085802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,7168,0.014313777287801107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,10240,0.12222933769226074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,3072,0.04698488778538174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,32,0.015499555402331881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,12288,0.050844444168938525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,16384,0.0951742198732164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,3072,0.01703288820054796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,16384,0.03623822331428528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,6144,2.9294266170925565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,8192,0.02010311186313629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,5120,0.6506880124409994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,3584,0.012839111189047495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,10240,0.021540444758203294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,2560,0.008393777741326226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,2560,0.06441955433951484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,8192,0.0942231085565355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,8192,0.3541902171240912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,10240,0.013401777380042605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,64,0.017532444662517972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,7168,0.015764445066452026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,6144,0.3358959886762831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,8192,0.4242053296830919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,256,0.005762666463851929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,3072,0.07445333401362102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,10240,0.03615999884075589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,4096,0.09807288646697998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,1536,0.02056711084312863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,3072,0.3187804487016466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,3584,0.16090933481852213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,256,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,64,0.0059279998143513995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,128,0.006027555714050929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,8192,0.018075555562973022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,3584,0.03942755526966519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,256,0.04133866561783685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,4096,0.014608888162506951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,8192,0.030689779255125258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,1536,0.11604444185892741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,3584,0.008590222232871586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,2560,1.842283460828993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,5120,2.1127707163492837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,7168,0.22609599431355795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,2048,0.008036444584528605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,512,0.00610133343272739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,6144,0.031122667921913996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,1024,0.006917333437336816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,768,0.007264888948864407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,128,0.004871110949251386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,2048,0.016367110941145156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,6144,0.03205688794453939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,2048,0.00612088872326745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,7168,0.014188443620999655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,2560,0.19651467270321318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,8192,0.033526221911112465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,51200,1.1307137807210286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,65536,0.25806932979159886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,512,0.004221333397759331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,12288,0.019524445136388142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,51200,0.18807289335462782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,7168,0.08650844626956516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,8192,0.11203200287289089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,5120,2.9148320092095266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,4096,0.017621333400408428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,1024,0.004127999974621667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,6144,0.01279022213485506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,3584,0.007857778006129796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,3072,0.009935110807418823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,1024,0.014707555373509726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,256,0.010978666444619497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,6144,0.03600800037384033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,65536,0.6899875534905328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,32,0.006717333363162146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,32,0.029806223180558946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,7168,4.301959991455078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,32,0.006734222173690796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,65536,0.06243911054399279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,4096,0.012936000194814471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,256,0.0037395556767781577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,512,0.00425866660144594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,32,0.0034213333080212274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,4096,0.05381777882575989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,1024,0.007088888850477006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,3072,0.01444355481200748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,512,0.06415022081798978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,8192,0.014904888139830695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,8192,0.02216177847650316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,1024,0.012511110968059964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,8192,0.0185244447655148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,65536,0.2234097851647271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,2560,0.06962222523159452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,32,0.00291377781993813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,1536,4.699622260199653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,10240,0.0217866665787167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,1536,0.01789955629242791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,8192,0.061166220241122775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,16384,1.1361138025919597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,6144,0.2052666611141629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,12288,0.12744266457027859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,2048,0.01311288856797748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,3072,0.008513777620262569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,768,0.005668444352017508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,10240,0.1863173378838433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,4096,0.019076445036464267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,256,0.004407999830113517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,8192,0.18920000394185385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,256,0.19592000378502739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,3072,0.01754488878779941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,16384,0.04887466629346212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,3072,0.01904266741540697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,64,0.015976000163290236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,1024,0.038633777035607234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,2048,0.0188417782386144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,3072,0.9087919659084744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,512,0.020872889293564692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,65536,0.729272895389133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,65536,0.09529244237475926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,5120,0.01035733355416192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,2048,0.006520000182920032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,256,0.10149777597851223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,6144,0.013463999662134381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,2560,0.08278844753901164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,1024,0.03295110993915134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,64,0.02970400121476915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,2560,0.10438666741053264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,256,0.01330044451687071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,256,0.0031600000543726813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,10240,0.006684444430801604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,6144,0.10438311100006104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,12288,0.1736640002992418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,512,0.016527111331621807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,768,0.007496000164084964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,2048,0.009534222384293875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,32,0.0027964444210131965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,3584,0.018613333503405254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,64,0.009993777506881291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,2048,0.0063288890653186375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,128,0.005207111024194294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,1024,0.0063822223908371395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,2560,0.05356088942951626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,8192,2.8645021650526257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,1024,0.006505777852402792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,8192,0.02280355493227641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,6144,0.05413066678576999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,32,0.007099555598364935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,2560,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,128,0.13370933797624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,128,0.006657777975002925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,3072,0.021594666772418555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,16384,0.05782933367623223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,128,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,1024,0.05244177911016676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,2560,0.016732444365819294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,2560,0.008921778036488427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,16384,0.023042667243215773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,32,0.005542222410440445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,1024,0.0058373332851462895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,128,0.005019555489222209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,768,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,3584,0.009117333425415887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,10240,0.09314311212963527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,16384,0.17260355419582793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,2048,0.021876444419225056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,3584,0.012844444149070315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,4096,0.008224888808197444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,256,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,1536,0.0092604441775216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,2560,0.03646488984425863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,768,0.007069333559936947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,3584,0.016117332710160148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,1024,0.6720968882242838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,5120,0.8969155417548286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,12288,1.9060221778021917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,256,0.012602667013804117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,512,0.0045422220395671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,10240,0.7275502416822645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,128,0.009719111025333405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,65536,0.4236355622609456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,768,0.04153866569201151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,1536,0.0061928890645504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,1024,0.006940444310506185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,3584,0.007914666500356462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,8192,0.2395084434085422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,64,0.02773155603143904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,12288,0.16656443807813856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,128,0.00721155520942476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,3584,0.01790488925245073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,3584,0.10748889048894246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,3584,0.017988445030318368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,10240,0.20705066786872017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,128,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,16384,0.7947484652201334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,768,0.08153244521882799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,7168,0.08538933595021565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,1024,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,256,0.45977242787679035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,10240,0.6372044351365831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,3584,0.13344000445471868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,1536,0.01019022199842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,12288,0.024298666252030268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,768,0.008429333567619324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,256,0.01536977787812551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,5120,0.018176888426144917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,8192,0.03439999951256646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,768,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,8192,0.04881422387229072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,6144,0.09479910797542995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,5120,0.02404977712366316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,512,0.007287999822033777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,2560,0.19779733816782633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,64,0.0028373334142896864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,3072,0.009455111291673448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,8192,0.06551200151443481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,4096,0.055049777030944824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,8192,0.020169778002632987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,12288,0.2569484445783827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,8192,0.17883643839094374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,64,0.00618488879667388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,1024,0.07714311281840007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,65536,0.07205244567659166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,1024,0.03632622294955783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,5120,0.011214222345087262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,2048,0.006333333336644703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,768,0.0035777779089079965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,2048,0.03654399845335219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,3072,0.08150577545166016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,1536,0.004648000001907349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,16384,0.09040355682373047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,6144,0.05726133452521431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,2560,0.09942755434248184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,16384,1.7228462431165907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,5120,0.019080888893869188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,64,0.008736000292830998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,2560,0.011536888778209686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,2048,0.01200711147652732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,128,0.042487111356523305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,2560,0.011206222077210745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,5120,0.01585422290696038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,768,0.019867555962668527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,12288,0.006070222291681502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,8192,0.03493511014514499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,10240,3.8185848659939237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,6144,0.01607466737429301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,1536,0.022416889667510986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,6144,0.01107377807299296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,256,0.005594666633341048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,1024,0.006156444549560547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,3584,0.010962666736708747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,12288,0.028078221612506445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,512,0.006055999961164262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,2048,0.12239644262525771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,256,0.017056889004177518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,4096,1.9496523539225261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,32,0.003371555565132035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,512,0.013544888959990608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,1536,0.48943201700846356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,7168,2.0417679680718317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,4096,0.024812445044517517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,32,0.005993777679072485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,2560,0.012599111431174807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,8192,0.029093331760830347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,16384,0.22869155142042372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,7168,0.015599111715952555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,2560,0.04883822136455112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,2048,0.07033422258165148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,1536,0.025685333543353613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,2048,0.7439449098375109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,128,0.10874399873945449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,16384,0.24865243169996473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,4096,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,4096,0.015455111861228943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,7168,0.011305777562989129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,4096,0.3711208767361111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,64,0.006226666685607698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,4096,0.18913600179884169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,8192,0.011647111011876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,4096,0.010216000179449717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,16384,0.021811554829279583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,128,0.0333315564526452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,7168,0.010888888604111142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,6144,0.25942667325337726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,65536,2.2060879601372614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,1536,0.012793777717484368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,3584,0.016514667206340365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,51200,0.7044640117221408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,2560,0.046498666206995644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,2048,0.09027200274997288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,3584,0.012156444291273752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,1024,0.04186044467820061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,1024,0.013648000028398303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,1024,0.020966221888860066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,5120,0.014179555906189812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,256,0.005544888890451855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,12288,0.0469359987311893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,2048,0.04489511251449585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,2560,0.008150222400824228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,65536,0.06612888971964519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,3072,0.04357599880960253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,6144,0.013930666777822705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,8192,3.843133290608724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,16384,7.912376403808594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,5120,0.010669333239396414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,65536,0.08394044637680054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,512,0.014896889527638754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,6144,0.10860800080829197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,1024,0.025199999411900837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,10240,0.18273066149817574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,1024,0.20133777459462485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,12288,0.027287999788920086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,65536,0.25009332762824166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,7168,0.01421333352724711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,2560,0.29439645343356663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,512,0.010743111371994019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,768,0.04648800028695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,2048,0.9990569220648872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,16384,0.2866151067945692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,3584,0.04654399885071648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,16384,0.12928088506062826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,51200,0.12459554937150742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,51200,0.10898933145735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,512,0.020115555988417733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,16384,0.029616001579496596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,768,0.0037991110649373797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,16384,0.043975111511018544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,3584,0.2685075600941976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,51200,0.792069329155816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,12288,0.08016977707544963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,64,0.005975111077229182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,64,0.014325333966149224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,3072,0.055504000849194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,51200,0.6283520062764486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,7168,0.01517688896920946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,51200,1.6068924797905817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,3584,0.01718844473361969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,4096,0.01574844453069899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,1536,0.0064195555945237475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,768,0.00444533344772127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,51200,0.10453066560957168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,2048,0.006775110960006714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,512,0.018087110585636564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,768,0.00721866637468338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,2048,0.07339377535714044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,12288,0.006036444256703059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,16384,0.014821334017647637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,1024,0.045145776536729604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,1024,0.011997333003415002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,3584,0.00810222244924969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,6144,0.07394399907853869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,6144,0.1283982197443644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,6144,0.05196266704135471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,16384,0.03415822320514255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,3584,0.01976444489426083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,2048,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,65536,0.14926932917700872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,2560,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,16384,0.6669626765780978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,1024,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,51200,0.08407200045055813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,32,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,7168,0.057020445664723717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,7168,1.992581261528863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,3072,0.021922666165563796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,768,0.00702133360836241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,1024,0.009107555780145857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,6144,0.2473804420895047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,1536,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,8192,0.007223111059930589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,16384,0.2679724428388807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,3072,0.015800888339678448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,6144,0.10616978009541829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,64,0.012643555800120035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,64,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,1024,0.3004071182674832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,128,0.004423111263248655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,7168,0.18764266702863905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,10240,0.09593155648973252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,4096,0.012968888713253869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,6144,0.009276444713274637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,1536,0.04282222191492716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,51200,0.3465564515855577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,768,0.0037644443412621817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,51200,1.9157191382514105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,51200,1.452112939622667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,6144,0.22047821680704752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,128,0.007473777565691207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,768,0.0041351111398802865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,512,0.05631999837027656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,8192,0.0507279998726315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,768,0.025467554728190105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,4096,0.02142133315404256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,3072,0.13027466668023005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,65536,0.10480355554156834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,2560,0.051649779081344604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,8192,0.032348444064458214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,4096,0.03226044442918565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,256,0.01603822244538201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,1536,0.012089778151777057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,6144,0.012959111067983838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,3584,0.14115199777815077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,6144,0.024895111719767254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,8192,0.08361066712273492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,5120,0.01014755583471722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,768,0.006335999816656113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,3584,0.021877333521842957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,32,0.0030702222138643265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,12288,0.27350221739874947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,10240,0.26508887608846027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,3072,0.008344000412358178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,10240,0.013597333596812354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,32,0.013776889277829064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,3584,0.010199111368921068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,3584,0.0247662216424942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,5120,0.009345778160625035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,10240,0.1639253298441569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,16384,0.2110524442460802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,512,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,64,0.009680888719028896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,51200,0.015511999527613321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,256,0.06461689207288954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,65536,4.212522718641493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,128,0.16428089141845703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,256,0.0052879999081293745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,512,0.007328000333574083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,128,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,4096,0.09624799754884507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,51200,0.21008711391025117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,32,0.0123848890264829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,8192,0.3602782090504964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,3584,0.01850400037235684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,51200,0.5174293518066406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,64,0.017078222499953378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,3584,0.06900711192025079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,6144,0.1356231106652154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,2560,0.021369778447681006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,7168,0.08907200230492486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,12288,0.015974221958054435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,51200,1.893238279554579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,512,0.007701333198282454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,1536,0.0075191110372543335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,768,0.19786577754550508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,10240,0.02492977844344245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,12288,0.03125155634350247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,32,0.030209776428010728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,256,0.0047422221137417685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,65536,0.285263114505344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,768,0.020830222302012973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,4096,0.007127111156781514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,65536,0.15997421741485596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,3072,0.014563555518786112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,256,0.0377111103799608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,10240,0.039204445150163435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,65536,0.31224176618787974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,12288,0.04071999920739068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,4096,0.3655129008822971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,1024,0.024519999821980793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,51200,1.5219039916992188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,256,0.0037173334923055437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,256,0.014696889453464083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,12288,0.09101421965493096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,10240,0.04088622331619263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,32,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,2048,0.02583644456333584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,6144,0.012409778104888068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,1024,0.021527111530303955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,12288,0.14020444287194148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,1024,0.24375287691752115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,51200,0.27781687842475045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,32,0.008929777476522658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,128,0.005888000130653381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,128,0.005443555613358815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,7168,0.013762666947311826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,2560,0.05968355470233493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,6144,0.01684888866212633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,128,0.004471999903519948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,8192,0.04560888806978861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,65536,1.7756338119506836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,1024,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,1024,0.004046222401989831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,16384,0.04340888725386726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,12288,0.018894222047593858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,768,0.005193777796294955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,7168,0.0710444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,12288,0.06082044707404243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,16384,0.02611911131276025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,4096,0.008408000071843466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,1536,0.051086223787731595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,512,0.08197333415349324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,10240,0.028007109959920246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,512,0.012703110774358114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,3584,0.01145066652033064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,64,0.02593155536386702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,8192,0.26254132058885365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,3584,0.006816000160243776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,8192,0.0898088879055447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,12288,0.22912356588575575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,2560,0.008999110923873054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,256,0.009462222456932068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,12288,0.3642862108018663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,51200,0.19539733727773032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,51200,0.05641511413786146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,7168,0.1506604486041599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,12288,0.21261600653330484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,51200,0.02499466637770335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,768,0.024784000383483037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,1536,0.10415733522839016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,65536,0.738425784640842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,51200,0.12095110946231419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,2048,0.009897777603732215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,5120,0.009466666314336989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,1536,0.02460622290770213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,128,0.005543111099137201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,256,0.0033075554917256036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,51200,0.16434666845533583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,768,0.02887822190920512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,12288,0.10688977771335179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,12288,1.239150259229872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,768,0.004860444615284602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,64,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,256,0.02936799989806281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,256,0.007088888850477006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,1536,0.01812977757718828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,256,0.005471999860472149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,1024,0.006436444405052397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,2560,0.014001778430408902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,10240,0.19492622216542563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,16384,0.08849244647555882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,256,0.002967111145456632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,65536,0.04543200135231018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,12288,0.04833955566088358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,16384,0.020290666156344943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,4096,0.07881689071655273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,512,0.008414222134484185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,64,0.01711466742886437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,16384,0.05448977814780342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,12288,0.18219289514753553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,7168,20.082544962565105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,4096,0.012867555850081973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,128,0.003576888806290097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,1536,0.03170488940344916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,65536,0.1626888910929362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,512,0.02202577723397149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,2048,0.13394221994611952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,256,0.016188444362746347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,3072,0.017685333887736004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,7168,0.05236000153753492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,65536,0.13707821898990208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,4096,0.006884444504976273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,1024,0.006072888771692912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,8192,0.21373155381944445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,5120,0.02628177735540602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,51200,0.05405333307054308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,64,0.011968889170222811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,7168,0.32649599181281197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,1024,0.01458666721979777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,6144,0.01501066651609209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,16384,0.10652799738778009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,512,0.014480888843536377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,2048,0.02271555529700385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,5120,0.011376889215575324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,1024,0.022293332550260756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,128,0.0028124445428450904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,512,0.006095110956165526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,5120,0.10175644689136082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,1536,0.02345688806639777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,5120,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,256,0.013243555194801755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,512,0.02645688917901781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,64,0.002540444541308615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,128,0.035143110487196184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,10240,0.030593779351976182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,65536,0.10136000315348308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,10240,0.3523413340250651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,4096,21.140229966905384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,4096,0.3258497714996338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,512,0.00702044450574451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,7168,0.11987822585635716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,5120,3.0596711900499134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,6144,0.019467555814319186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,51200,0.3305911223093669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,512,0.008408889174461365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,128,0.015450666348139444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,1024,0.09638044569227432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,16384,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,3072,0.07285955879423353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,768,0.04538755615552267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,10240,0.1895600027508206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,1536,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,65536,5.141657935248481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,2048,0.010500444306267632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,1536,0.007370666497283512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,2560,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,16384,0.3242275449964735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,2048,0.038183109627829656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,2560,0.008048888709810045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,3072,0.012695111334323883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,5120,0.1953057712978787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,32,0.025825776987605627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,51200,0.09634844462076823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,3072,0.02092622220516205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,10240,0.030594666798909504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,128,0.00272088891102208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,3584,0.01092088884777493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,7168,0.42591378423902726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,1536,0.026351110802756414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,10240,0.22146756119198271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,8192,0.02862488892343309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,64,0.006667555620272954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,8192,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,7168,0.016365332735909354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,3584,0.02882844540807936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,3584,0.0964426663186815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,6144,0.028685334655973647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,1536,0.460593753390842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,2560,0.012124444047609964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,1536,0.010619555910428366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,768,0.2698266771104601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,768,0.006783111227883234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,256,0.0032711111836963226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,16384,0.1241777737935384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,4096,0.09722310966915554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,12288,1.5835030873616536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,3072,0.05353066656324598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,10240,0.01664355562792884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,3584,0.0068915556702348925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,4096,0.014695111248228284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,1536,0.004670222186379963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,7168,0.17362756199306914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,5120,0.016081778539551627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,51200,0.41106756528218585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,2048,0.026484444737434387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,1024,0.012640888492266336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,1024,0.01349244432316886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,32,0.002748444469438659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,256,0.002871111035346985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,10240,0.05877066983116997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,1536,0.006367110957702001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,65536,0.022802667485343084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,3584,0.1653404500749376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,64,0.002748444469438659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,8192,0.01314933349688848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,3584,0.3524213367038303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,2560,0.8793146875169543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,1024,0.04555288950602213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,1024,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,2048,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,7168,0.015470221638679504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,12288,0.0464897784921858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,256,0.007296889192528195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,32,0.007320000065697565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,256,0.009034666750166152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,3072,0.012321777641773224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,6144,0.015824889143308003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,256,0.1969511111577352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,4096,0.045773333973354764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,10240,0.01903733279969957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,8192,0.06431733237372504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,128,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,6144,0.1483502255545722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,2048,0.02775466607676612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,512,0.02709688908523983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,2560,0.011090666883521609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,512,0.015126221709781222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,1024,0.006892444358931647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,8192,0.05048977666431003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,16384,1.4369867112901475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,16384,0.056806219948662647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,1024,0.010503110786279043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,3584,0.044143110513687134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,2560,0.015103111664454142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,3072,0.15172621938917372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,5120,0.02311377724011739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,32,0.007326222128338284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,2560,0.026687999566396076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,2560,0.022944000032213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,5120,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,16384,0.5480088657803005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,128,0.010461333725187512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,12288,0.2528657648298475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,4096,0.1877839962641398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,16384,1.5407644907633464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,64,0.0047084444926844705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,512,0.004387555436955558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,4096,0.06151022513707479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,5120,0.016120889120631747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,3584,0.011287111375066968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,512,0.0047386665311124595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,64,0.005680000202523337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,51200,0.05925599733988444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,65536,0.013569778038395775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,7168,0.021595555875036452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,4096,0.02385955552260081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,128,0.006257777826653586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,256,0.008295111358165741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,65536,0.022356443934970435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,512,0.004868444469239977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,8192,0.020374221934212577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,768,0.0037946667936113146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,16384,0.01975288821591271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,3584,0.019523556033770244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,10240,2.3643902672661676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,4096,0.04795555604828728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,1024,0.006439111298984951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,12288,0.21018400457170275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,2048,0.018467555443445843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,1024,0.006508444332414203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,4096,0.00905511114332411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,5120,0.01148622234662374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,51200,1.0099884668986003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,1536,0.005534222142563925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,128,0.013373333546850415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,128,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,1024,0.006735111276308696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,3584,0.10192977719836765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,3584,0.025204444924990337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,12288,0.01683111157682207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,6144,0.02991466720898946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,32,0.050719112157821655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,1024,0.005224888937340842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,3584,0.05107644531461927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,51200,0.1963813304901123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,128,0.004062222109900581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,3584,0.014521777629852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,2560,0.03354933195643955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,10240,0.01604711181587643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,2560,0.01516533394654592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,1536,0.008618666893906063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,8192,0.02036088870631324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,2560,0.04828888840145535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,512,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,4096,0.01753866672515869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,512,0.35872001118130153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,1024,0.04708088768853081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,128,0.016923555069499545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,12288,0.5370533201429579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,12288,0.293757332695855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,5120,0.1669715510474311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,2048,0.005183111048407025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,1536,0.011071999867757162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,32,0.006109333286682765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,4096,0.019732443822754752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,7168,0.01800355480776893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,6144,0.3312995433807373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,3072,0.01734133395883772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,12288,0.18783111042446562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,10240,0.02163377735349867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,1536,0.05097155438529121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,65536,0.14221155643463135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,6144,0.03715199894375271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,5120,0.015172445111804538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,2048,0.08828977743784587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,3584,0.05411200059784783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,512,0.04807466599676344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,4096,0.012702222499582501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,12288,0.023989333046807185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,3584,0.13300622834099665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,128,0.00829777783817715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,8192,0.011889778077602386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,6144,0.017772444420390658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,1024,0.014486221803559197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,6144,0.11289332972632514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,1536,0.007010666860474481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,8192,0.2952444553375244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,16384,0.060200889905293785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,8192,0.03592000073856778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,3584,0.008221333225568136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,5120,0.22203732861412895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,1024,0.008225777910815345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,12288,0.19043555524614122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,3584,0.14177600542704263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,128,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,51200,0.17331733968522814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,65536,0.06183555391099718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,10240,0.027143999934196472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,10240,0.11577778392367893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,7168,0.17382400565677217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,8192,6.113470289442275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,3584,0.20904178089565703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,2560,0.05490666627883911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,3072,0.008912000391218398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,512,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,512,0.023964444796244305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,6144,0.011715555356608497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,16384,0.5608168707953559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,256,0.2438488801320394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,1024,0.014422222971916199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,2048,0.007153777612580194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,5120,0.07963911030027601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,2048,0.012002666791280111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,5120,0.009951111343171861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,10240,0.20802933639950225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,4096,0.008295111358165741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,51200,5.278271993001302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,512,0.017824888229370117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,8192,1.900616963704427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,8192,0.8459999826219348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,256,0.016527111331621807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,4096,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,3072,0.09862311018837823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,32,0.003636444608370463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,5120,0.6899058024088541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,16384,0.23643289672003853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,128,0.005240889059172736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,768,0.004808889081080754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,16384,0.006673777682913675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,2560,0.017847999930381775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,768,0.015034667319721647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,3584,0.045810666349199086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,3072,0.013376000026861826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,8192,1.6611582438151042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,1024,0.007697777615653143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,4096,0.011429333024554782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,2560,0.009492444495360056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,51200,0.45106665293375653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,3584,0.015777778294351365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,12288,0.015359110302395292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,1024,0.008444444172912175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,3584,0.12515555487738714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,16384,1.2755493587917752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,12288,0.008811555802822113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,5120,0.011782222323947482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,1536,0.008715555899673039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,32,0.006084444622198741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,3584,0.03457155492570665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,10240,0.11339378356933594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,3584,0.016138666205936007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,32,0.010832889212502373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,128,0.007804444266690149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,1536,0.05905422237184313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,1536,0.015235554840829639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,65536,2.06368531121148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,64,0.003383999897374047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,4096,0.011743110915025076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,16384,0.08337155315611099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,3584,0.013244444297419654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,8192,0.015410666664441427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,5120,0.04692799846331278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,8192,0.014680888917711047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,1024,0.005020444591840108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,7168,0.027648889356189307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,51200,0.061662223603990346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,512,0.09011022249857585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,16384,0.034352001216676496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,10240,0.15160977840423584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,2560,0.012915555801656512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,8192,0.20100888941023085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,51200,0.0600186652607388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,256,0.020875554945733812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,6144,0.012869333227475485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,4096,0.04426133301522997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,51200,23.656656901041668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,10240,0.019465777609083388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,512,0.017708443933063082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,12288,0.1564026673634847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,128,0.05537244346406725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,2560,0.06956266694598727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,51200,33.022013346354164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,1536,0.05064622229999966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,2048,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,4096,0.03576888971858554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,2048,0.007905777957704332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,10240,0.04120444589191013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,512,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,1024,0.03420799970626831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,5120,0.1253395610385471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,768,0.06988710827297635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,3072,0.030753778086768255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,5120,0.13842399915059408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,256,0.003061333257291052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,65536,0.0685173339313931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,2048,0.3151155577765571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,3072,0.007553777760929531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,12288,0.17530222733815512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,3584,0.017450667089886136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,65536,0.052926222483317055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,65536,0.16956267091963026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,5120,0.031130668189790513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,64,0.006544000158707301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,7168,0.06068266762627495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,7168,0.05146933264202542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,512,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,1024,0.012890666723251343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,128,0.006741333338949416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,3584,0.19883555836147734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,2560,0.04703022374047173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,8192,0.324127991994222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,32,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,8192,0.0632293356789483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,1536,0.025663110944959853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,12288,1.2609003914727104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,32,0.028051555156707764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,8192,0.34560712178548175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,6144,0.007887110941939885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,2560,0.006320888797442119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,1536,0.013515555196338229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,7168,0.0236124445994695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,1536,0.00463733325401942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,5120,0.022898667388492163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,64,0.009714666340086196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,512,0.004971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,64,0.011619555453459421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,10240,0.5648729006449381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,12288,0.020523554748959012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,4096,0.05986844168768989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,1536,0.06842933098475139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,10240,0.057137780719333224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,65536,0.10092888938056098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,51200,0.09361600213580662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,10240,0.1511173380745782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,4096,0.010450666977299584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,768,0.005412444472312927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,8192,0.26201510429382324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,5120,0.015768888923856948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,1536,0.006264888991912206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,65536,0.526135974460178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,1024,0.018170666363504197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,4096,0.012422222230169507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,10240,0.016013332539134555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,2048,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,2048,0.014669333895047506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,7168,0.05683022075229221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,3584,0.07725155353546143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,4096,0.17816266748640272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,6144,0.1122631099489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,4096,0.01089599976936976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,65536,0.6626764403449165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,12288,0.0070266665683852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,32,0.007784888976150089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,256,0.003571555432346132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,65536,0.29374755753411186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,51200,0.06390844451056586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,12288,0.04485244552294413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,128,0.03160711129506429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,128,0.004597333156400257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,4096,0.08919822507434422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,5120,0.05685333410898844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,10240,0.33392977714538574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,7168,0.010543111297819348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,2560,0.006333333336644703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,3072,0.10829244719611274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,51200,0.8441804250081381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,768,0.005018666800525454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,16384,0.057224002149369985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,256,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,2560,0.11253155602349176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,512,0.0075431110130416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,51200,0.08294666475719877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,7168,0.14453155464596218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,256,0.008601778083377415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,512,0.07625866598553128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,768,0.004381333374314838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,51200,0.09656177626715766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,16384,0.01521511044767168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,10240,0.8283697764078776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,7168,0.18697422080569795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,2048,0.015095111396577625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,16384,0.025991999440722998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,6144,0.07733600007163154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,51200,1.7592444949679906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,2560,0.48701153861151797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,1024,0.006723555425802867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,51200,1.6229226854112413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,51200,0.09738310840394761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,4096,0.009776888622177972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,128,0.011069333387745751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,128,0.0041928887367248535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,4096,0.017465778523021273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,128,0.005324444423119227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,10240,0.25081690152486164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,64,0.005505777895450592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,6144,1.2386870914035373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,65536,0.14934933185577393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,10240,2.11280886332194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,4096,0.10522488753000896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,5120,2.0785768296983504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,51200,0.07107910845014784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,128,0.017614222235149808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,6144,0.01068800025516086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,2560,0.220378663804796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,7168,0.015302222636010913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,4096,0.011794666449228922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,2048,0.006555555595291986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,51200,0.9962097803751627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,8192,0.1707866721683078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,2560,0.016188444362746347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,51200,0.1034711135758294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,7168,0.00847111145655314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,4096,0.01421511173248291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,3072,0.04767733481195238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,12288,0.09988800022337173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,768,0.003648888733651903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,4096,0.013491555220550962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,12288,0.019694222344292533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,10240,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,3072,0.010483555495738983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,7168,0.10214933421876694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,256,0.004632888982693355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,12288,0.006203555398517185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,51200,0.29135823249816895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,51200,3.497305764092339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,5120,0.15224266052246094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,128,0.004819555415047539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,1536,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,64,0.012869333227475485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,1024,0.03908355699645148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,768,0.003943110919660992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,8192,0.018040888839297824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,7168,0.011752888560295105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,32,0.04407377706633674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,10240,2.1417325337727866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,3072,0.012230222423871359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,1536,0.03366933266321818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,3072,0.059915555848015674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,768,0.004743111216359668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,12288,0.3235022226969401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,2048,0.006056889063782162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,2048,0.6491653124491373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,4096,0.012958221965365939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,10240,0.3778995672861735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,6144,0.023500444160567388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,3584,0.01613066593805949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,32,0.008769777913888296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,10240,0.1993795500861274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,3584,0.19556266731686064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,1024,0.007842666572994655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,4096,0.010678222609890832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,1024,0.006375999914275275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,65536,0.18943377335866293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,3584,0.01832355558872223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,16384,0.02019644445843167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,768,0.004855111241340637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,768,0.021398221453030903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,6144,0.00794044468137953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,3584,0.026914666096369427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,4096,0.012479999827014076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,16384,0.4923217561509874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,1024,0.014328888720936246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,1536,0.303783999549018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,768,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,65536,0.06680355469385783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,65536,0.1729760037528144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,1536,0.007108444141017065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,7168,0.1411493354373508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,65536,0.17098132769266763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,3584,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,128,0.011200000014570026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,768,0.06541599829991658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,4096,0.00906399968597624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,1536,0.13496355215708414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,2560,0.18894133302900526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,10240,0.01032266683048672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,3072,0.027168888184759352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,7168,0.11672622627682155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,256,0.003056888985964987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,51200,0.07607111003663805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,1536,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,7168,0.31444533665974933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,12288,0.17790755960676405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,3072,0.006586666736337874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,6144,0.10624177588356866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,8192,0.8405600123935275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,7168,0.07867022355397542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,12288,0.015198222464985318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,12288,0.09442133373684353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,2048,0.06618133518430921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,3072,0.007483555210961237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,65536,0.03915289044380188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,2560,0.05814044343100654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,256,0.00535111129283905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,512,0.022946667340066697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,2048,0.13846755027770996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,51200,0.3615893257988824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,10240,0.2600444422827827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,5120,0.01920088960064782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,256,0.007177777588367462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,10240,0.5522106488545736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,768,0.005254222287072076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,65536,0.08738133642408584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,512,0.022499556342760723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,2048,0.024137778414620295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,3072,0.012574222352769641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,6144,0.0284826656182607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,64,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,1024,0.02420355545149909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,512,0.2265911102294922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,10240,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,10240,0.10406311353047688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,12288,0.0431600014368693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,1024,0.29687378141615123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,3584,0.012050666742854647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,10240,0.17082577281528047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,512,0.02273066673013899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,2560,0.007672888537247975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,2048,0.13767200046115452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,512,0.03355288836691115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,768,0.005903999838564131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,2560,0.020984888076782227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,256,0.004341333276695675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,128,0.005406222409672207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,5120,0.25395554966396755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,128,0.004479111068778568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,5120,0.2429306772020128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,512,0.14815555678473577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,2560,0.023600889576805964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,12288,1.4729493459065754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,512,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,1536,0.0337413317627377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,65536,0.38143377833896214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,3584,0.014459555347760519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,3584,0.7190497716267904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,1024,0.14339021841684976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,512,0.007620444728268518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,768,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,3072,0.012871111432711283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,10240,0.007171555525726742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,2048,0.04210577739609612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,7168,0.2565404574076335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,8192,0.08435022168689305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,16384,0.06083466609319051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,32,0.007517333659860823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,8192,0.13355555799272326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,2048,0.1990248892042372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,32,0.00867733359336853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,5120,0.12746222813924155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,5120,0.008449777960777283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,10240,0.018812444474962022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,1024,0.005734222216738596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,512,0.006575999988449945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,4096,0.012903999951150684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,65536,0.3759875562455919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,51200,1.9837936825222438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,65536,0.16180888811747232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,32,0.025478222303920325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,32,0.005030222237110138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,10240,0.02398666739463806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,3584,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,1536,0.008040000167157915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,2048,0.6948417557610406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,8192,0.015579554769727917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,65536,0.10232088963190715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,64,0.003047111133734385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,8192,0.015117333994971381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,16384,0.022983999715911016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,10240,0.046224001381132335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,2048,0.06590755780537923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,1024,0.6291919814215766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,1024,0.01690933273898231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,12288,0.28316712379455566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,256,0.12689956029256186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,3584,0.03577155537075467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,1536,0.11873866452111138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,16384,0.01928711103068458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,5120,0.012336889074908363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,16384,0.47281424204508465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,65536,0.39780354499816895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,16384,0.35845422744750977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,10240,0.07413511143790351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,7168,2.595066706339518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,12288,0.4763306511773004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,3584,0.13587289386325413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,128,0.005041777673694822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,5120,0.012915555801656512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,4096,0.011316444310877057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,5120,0.06351733207702637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,5120,0.02003466586271922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,32,0.0037128888070583344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,2560,0.009296889106432596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,1024,0.09941689173380534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,512,0.18310756153530547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,16384,0.0565333366394043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,5120,0.04460266563627455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,12288,0.00923999978436364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,2048,0.2424080106947157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,128,0.04205066627926297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,10240,0.36247022946675617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,3072,0.06289244360393949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,10240,0.01516177753607432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,2560,0.017342221405771043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,256,0.019786667492654588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,7168,0.015113777584499784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,1536,0.007368889119890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,768,0.008953777452309927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,51200,0.2437288761138916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,3584,0.029480000336964924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,6144,0.12065688769022624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,8192,0.018290667070282828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,3584,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,128,0.09034222364425659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,1536,1.039354642232259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,1536,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,768,0.009602666729026372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,8192,0.03955022162861294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,7168,0.7075040075514051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,1536,0.020663999848895602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,7168,0.01774133410718706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,6144,0.22436889012654623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,2560,0.10307021935780843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,5120,0.061223109563191734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,128,0.02027555637889438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,7168,0.047396444612079196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,1536,0.02456355591615041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,2048,0.07637866338094075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,6144,0.04399111204677158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,2048,0.005316444569163852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,256,0.008873778084913889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,1024,0.014488000008794995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,3584,0.08982133203082615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,3072,0.1225866609149509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,128,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,65536,2.1233893500434027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,7168,0.027816888358857896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,10240,4.005022260877821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,512,0.008755555583371056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,10240,0.1302328904469808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,32,0.00812266684240765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,4096,0.07322133249706693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,3072,0.036517331997553505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,128,0.018124444617165458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,8192,0.11724000506930882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,256,0.0034106667670938703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,6144,0.021366222037209406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,12288,0.015037332971890768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,1024,0.003979555434650845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,32,0.004150222159094281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,10240,0.44357866711086696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,6144,0.11066933472951253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,65536,0.45854753918117946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,768,0.005734222216738596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,1536,0.004930666751331753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,2048,0.012330667012267642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,10240,0.018404444058736164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,512,0.006909333169460297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,512,0.0031884445084465873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,2048,0.1782942215601603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,10240,0.0206657780541314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,2048,0.05851466788185967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,65536,19.96409437391493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,2560,0.07913866970274183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,10240,0.15563733047909206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,8192,3.678560045030382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,1024,0.005734222216738596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,256,0.005872888697518243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,3072,0.030061334371566772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,4096,0.10899022552702163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,51200,0.13304089175330266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,3072,0.024977778395016987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,256,0.0042408886882993905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,2560,0.020432000358899433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,32,0.02086399992307027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,1024,0.06405066781573825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,4096,0.011922666596041786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,256,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,5120,0.02548888822396596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,64,0.008961777720186446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,4096,0.04894222153557671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,64,0.002604444407754474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,6144,0.029820445511076186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,65536,0.9165795644124349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,256,0.1133413314819336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,10240,0.04804889029926724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,8192,0.1135279999838935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,8192,0.08458844158384536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,1024,0.06678844160503812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,2560,0.007689778175618913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,2048,0.006211555666393704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,64,0.14328000280592176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,8192,0.01976088848378923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,2048,0.012967111335860359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,128,0.003955555458863576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,7168,0.010585777461528778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,65536,0.030833777454164293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,3584,1.6352230707804363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,3584,0.008825778133339351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,2048,0.011360000405046674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,768,0.010682666467295753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,1024,0.04511911008093092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,51200,0.9886177910698785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,32,0.006617777877383762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,3584,1.4330853356255426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,4096,0.1431866619322035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,768,0.007325333025720384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,16384,3.0198578304714627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,2048,0.04416088925467598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,128,0.008426666259765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,2560,0.07846577962239583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,5120,0.010415111151006486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,512,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,5120,0.03752177622583177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,512,0.005756444401211209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,2560,0.007385777930418651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,512,0.006180444525347815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,3072,0.0844693316353692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,8192,0.018522666560279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,256,0.006693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,16384,5.17996088663737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,51200,0.05488444368044535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,4096,0.2515653239356147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,10240,0.026038222842746313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,128,0.015030221806632148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,6144,1.1875137752956815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,3072,0.009961778091059791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,3584,0.02424088948302799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,32,0.005763555566469829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,8192,0.05640266338984171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,128,0.0047031111187405055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,51200,0.6623857816060384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,128,0.005719111197524601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,512,0.003597333199448056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,1024,0.03701688845952352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,32,0.005165333135260476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,2560,0.009076444639099969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,7168,0.023285332653257582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,8192,0.02658577760060628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,2048,0.01239911135700014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,4096,0.00904088881280687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,3072,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,6144,0.014130666851997375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,5120,0.04639022217856514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,64,0.002521777732504739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,4096,0.013688888814714221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,3584,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,2048,0.15758577982584634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,2048,0.007909333540333642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,2048,0.019702222612169053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,2560,0.013345777988433838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,4096,0.2175439993540446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,65536,0.10431822141011556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,128,0.006913777854707506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,12288,0.7916702164543999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,1024,0.014556444353527494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,7168,0.02161066730817159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,10240,0.06061422162585788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,7168,0.011025778121418424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,128,0.00282577777074443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,1536,0.009890666438473595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,12288,2.9629688262939453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,12288,0.13587733109792074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,2560,0.10510400268766616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,12288,0.029807110627492268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,768,0.013019555144839816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,51200,0.09577688905927871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,3072,0.009955555200576782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,3584,0.01608088943693373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,65536,8.224713643391928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,256,0.0062871111763848206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,256,0.02144444485505422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,51200,0.3357813358306885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,51200,0.013841778039932251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,51200,0.06072355641259087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,1536,0.031769778993394636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,10240,0.017449777987268236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,4096,0.2332515451643202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,16384,0.4779591030544705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,6144,0.048136889934539795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,3072,0.03398666779200236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,512,0.016809778081046212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,65536,0.12350311544206406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,7168,0.061111112435658775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,128,0.004663111021121343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,512,0.023356444305843774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,51200,83.94204372829861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,64,0.005747555444637935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,51200,1.3556204901801214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,768,0.00962400022480223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,12288,0.02628266645802392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,768,0.01923111081123352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,3072,0.41899821493360734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,7168,0.01464000013139513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,1536,0.007169778148333232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,4096,0.02314577831162347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,256,0.02325155503220028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,128,0.01315111087428199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,32,0.037109331952200994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,2048,0.20285422272152373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,128,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,5120,0.009387555221716562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,3072,0.01870666609870063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,7168,0.14005066288842097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,1024,0.02795999911096361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,8192,0.115283555454678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,64,0.009935999910036722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,256,0.004755555755562252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,1024,0.029050668080647785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,256,0.033209777540630765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,3072,0.02424088948302799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,3072,0.008810666700204214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,4096,0.010288889209429422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,512,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,1536,0.03828533490498861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,10240,0.06779288583331637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,4096,0.026309332913822595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,7168,0.006423111177153057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,256,0.015802666544914246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,768,0.21831999884711373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,1536,0.010115555591053432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,512,0.006139555739031897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,3072,0.010781333678298526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,64,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,65536,6.549740261501736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,1536,0.03012266755104065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,10240,0.011152000062995486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,4096,0.0477804442246755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,768,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,768,0.038219554556740656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,51200,0.28599821196662056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,256,0.007921777665615082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,3584,0.06761244270536634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,65536,0.15674755308363172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,256,0.009740444521109263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,6144,0.014076444837782117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,8192,0.022851554883850947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,12288,0.3701155450608995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,64,0.08920088741514419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,7168,0.5734560224745009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,51200,0.20372533798217773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,256,0.004053333153327306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,8192,1.4453208711412218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,512,0.2256284554799398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,256,0.009485333330101436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,6144,0.05072355601522657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,128,0.15521600511338976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,5120,0.027245332797368366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,1536,0.07711733049816556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,5120,0.028563555743959215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,8192,0.09461955229441325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,6144,0.2071039941575792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,768,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,8192,0.026170666019121807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,12288,0.21557422478993735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,256,0.004848888764778773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,3072,0.01479022204875946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,6144,0.17762133810255262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,7168,0.01992177797688378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,16384,0.31082044707404244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,51200,0.15598666667938232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,768,0.0063386667105886675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,1536,0.015341333217091031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,6144,0.041182223293516375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,256,0.00980799976322386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,7168,0.029584887954923842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,3072,0.3477164374457465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,8192,0.07198222478230794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,32,0.0027235555979940626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,10240,0.3566017680697971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,51200,0.09151911073260838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,2048,0.005320000151793162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,3584,0.015081778168678284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,12288,0.020398222737842135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,51200,3.0745360056559243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,512,0.018434666925006442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,1536,0.008831111093362173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,32,0.007526222202512953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,65536,0.028892444239722356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,51200,3.3680311838785806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,1536,0.046053333414925464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,65536,0.40706488821241593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,3072,0.05156977640257942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,768,0.004489777816666497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,12288,0.23926311069064668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,16384,0.11195111274719238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,512,0.4543813069661458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,4096,0.02649066680007511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,3584,0.11624977323744033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,512,0.0064604443808396654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,2560,0.02019466625319587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,1536,0.19696533679962158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,512,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,16384,0.3190026548173693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,5120,0.40374755859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,64,0.005632888939645555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,64,0.021154666940371197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,3584,0.011617778076065911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,3072,0.07932088772455852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,1536,0.005028444445795483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,768,0.005817777580685086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,1536,0.01145066652033064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,3584,0.009827555881606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,3072,0.03422222203678555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,8192,0.21411911646525064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,3072,0.2837493419647217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,768,0.09002311362160577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,8192,0.3403502305348714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,12288,0.3588675657908122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,7168,0.021749332547187805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,16384,0.09009155299928452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,2048,0.046345776981777616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,7168,0.010992888775136737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,3072,0.07007733318540785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,8192,0.012216889195972018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,8192,0.016042666302786935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,32,0.002990222225586573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,512,0.014256000518798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,6144,0.01564800077014499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,16384,0.019497777024904888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,1024,0.00598933340774642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,10240,0.060191108120812305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,7168,0.01722755531469981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,16384,0.1023377776145935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,64,0.0026026666164398193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,12288,0.02828888760672675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,2560,0.006200888918505774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,5120,0.025096888343493145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,128,0.0040755557517210645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,4096,0.27046044667561847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,64,0.0026088888860411118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,51200,3.8660240173339844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,65536,0.00999644481473499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,6144,0.012591111163298288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,5120,0.012961777547995249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,3584,0.0703315536181132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,16384,0.006731555693679386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,7168,0.12901688946617976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,128,0.0033813334173626373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,512,0.137210660510593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,12288,0.018401778406567044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,16384,0.07061244381798638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,7168,0.8022942013210721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,10240,0.061871111392974854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,768,0.00814666681819492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,3072,0.02148799929353926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,5120,0.02976089053683811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,8192,0.16113600465986463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,7168,0.262043555577596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,3584,0.02509422269132402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,3584,0.01886577738655938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,64,0.005583999885453119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,8192,0.10144978099399143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,4096,0.0633102191819085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,256,0.005111999809741974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,65536,0.36493423249986434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,256,0.013227555486891003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,768,0.004304888761705822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,3072,0.6695413589477539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,10240,0.04382488793796963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,12288,0.018526222970750596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,1536,0.01110755569405026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,16384,0.34343645307752824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,512,0.024975111087163288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,16384,1.5370968712700739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,3584,0.014888000157144336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,2048,0.0057546666098965546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,65536,0.147070222430759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,12288,0.02107911143038008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,65536,0.6437395413716634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,32,0.03454400102297465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,1536,0.023507555325826008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,65536,0.08158222171995375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,65536,0.0400257772869534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,4096,0.020658666888872784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,51200,0.27200176980760365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,4096,0.012445333103338877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,3072,0.008749333520730337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,4096,0.3715395662519667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,128,0.005030222237110138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,512,0.0075662218862109715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,65536,0.2604897816975911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,8192,0.11339910825093587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,2048,0.019486222002241347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,6144,0.042436444097095065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,128,0.014087999860445658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,256,0.007831999825106727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,4096,0.005868444426192178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,65536,6.6428638034396705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,6144,0.01463822192615933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,12288,0.5587360064188639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,256,0.02057422200838725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,2560,0.34960177209642196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,1536,0.014180445008807711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,7168,0.19482221868303087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,4096,0.013880000346236758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,768,0.0090506664580769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,2560,0.01279022213485506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,64,0.08444711234834458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,64,0.0029484445436133277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,10240,0.02236088944805993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,2560,0.04868444469239977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,4096,0.03358577688535055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,51200,0.10425688823064168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,51200,0.14407466517554388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,2560,0.02587822245226966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,7168,0.019039111004935372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,2560,0.07249955336252849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,16384,0.136790222591824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,128,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,4096,0.02496622171666887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,2048,0.01643288963370853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,5120,0.0124177775449223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,8192,0.03232533401913113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,5120,0.025234666135576036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,256,0.015153777268197803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,2560,0.009360888765917884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,1536,0.010770666930410596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,2048,0.014224888549910652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,6144,0.01828977796766493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,5120,0.18492089377509224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,32,0.006590222318967183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,3584,0.011687111523416309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,4096,0.04551911022928026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,32,0.007624000310897827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,512,0.007552000383536021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,4096,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,65536,0.10569777753618027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,51200,0.10904177692201401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,3072,0.013869333598348828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,256,0.0403386652469635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,65536,0.0660968886481391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,32,0.002847111059559716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,65536,0.3851599958207872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,1536,0.0731217794948154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,10240,0.03698044353061252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,5120,0.018928888771269057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,12288,0.5808844566345215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,512,0.030596445004145306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,65536,0.9710968865288628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,256,0.003951999876234266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,256,0.004440000073777305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,128,0.003952888978852166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,8192,0.3461315631866455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,2560,0.008276444342401292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,1536,0.007459555235173967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,12288,0.03024622135692173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,512,0.0033288889875014624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,4096,0.0116462219092581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,51200,0.16558044486575657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,12288,0.01686311099264357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,12288,0.03228622343805101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,768,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,2560,0.01628533336851332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,16384,0.020118221640586853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,2048,0.01759466694460975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,3584,0.014183110660976835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,512,0.00443466669983334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,3072,0.11514311366611057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,64,0.03372177812788222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,512,0.007086222370465596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,512,0.004913777940803104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,256,0.02077244387732612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,512,0.021301334102948506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,512,0.005401777724424998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,5120,0.16279289457533094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,128,0.0027964444210131965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,51200,1.0130364100138347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,8192,0.42064444224039715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,2048,0.018225777480337355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,768,0.027107555005285475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,16384,0.07458311319351196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,1024,0.06345422400368585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,2048,0.009621333744790819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,5120,0.01864888932969835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,8192,0.06584799951977201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,768,0.006337777607970768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,128,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,4096,0.13807466295030382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,8192,0.39757421281602645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,6144,0.03293155630429586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,2048,0.008488000267081791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,5120,0.34764711062113446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,4096,0.30759112040201825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,6144,0.07185866435368855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,2560,0.03327466713057624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,2048,0.03163377775086297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,7168,0.16142222616407606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,16384,9.121564229329428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,768,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,256,0.018927110566033255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,2560,0.36740888489617246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,3584,0.01038755559258991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,1024,0.010291555689440833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,8192,0.08484444353315566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,51200,0.19293777147928873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,3584,0.022152889106008742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,7168,0.1129208935631646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,256,0.011115555961926779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,51200,1.2719519933064778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,3072,0.011448889142937131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,2048,0.0074337778819931885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,3072,0.021253334151373968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,512,0.008848889006508721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,10240,0.03054666519165039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,4096,0.013010666602187686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,1024,0.0069297779765393995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,5120,0.016149333781666227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,12288,0.0691608919037713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,128,0.0042071110672420925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,51200,0.05266666743490431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,1536,0.045040001471837364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,64,0.1931075519985623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,6144,0.028372443384594385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,6144,0.04129244552718268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,16384,0.4149973392486572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,1536,0.005632888939645555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,2560,0.00852888905339771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,256,0.043064001533720225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,3072,0.13941689332326254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,4096,0.011571555501884885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,10240,5.166778564453125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,256,0.004193777839342753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,512,0.007469333708286285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,512,0.030930668115615845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,4096,0.015394666128688388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,65536,0.4307991133795844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,51200,0.09267466598086888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,6144,0.0654675563176473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,1536,0.09361777702967326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,2560,0.009933333430025313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,5120,0.007484444313579136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,5120,0.021312000022994146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,7168,0.011763555308183035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,1536,0.018745778335465323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,2560,0.02938133478164673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,1536,0.006864000111818314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,2560,0.10066044330596924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,6144,0.016374222106403775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,1024,0.010178666975763109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,32,0.005514666852023866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,1024,0.00997511131895913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,1024,0.004044444610675176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,3584,0.008889777792824639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,51200,0.06081155273649427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,16384,0.13342311647203234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,8192,0.02072266737620036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,1536,0.004716444346639845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,10240,0.47855022218492294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,7168,0.01368711143732071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,3584,0.00793599999613232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,3584,1.965152104695638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,2048,0.011951111257076263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,3584,0.01332444449265798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,3072,0.047135998805363975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,65536,0.02495644489924113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,1024,0.006113777971929974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,256,0.007904888855086433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,51200,1.6357964409722223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,2048,0.02518133322397868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,2560,0.007181333170996771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,3584,0.00868000007337994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,2048,0.0058559998869895935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,1536,0.007308444215191736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,512,0.01001422190003925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,768,0.005538666827811136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,128,0.02385777731736501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,4096,0.2685733371310764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,256,0.018925334016482036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,256,0.03887733485963609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,2560,0.01830933325820499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,512,0.016208888755904306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,7168,0.13561067316267225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,10240,0.13021777735816106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,3072,0.010229333407349056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,8192,0.08880177471372817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,12288,0.4831804699367947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,10240,0.03152533372243246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,7168,0.051856001218159996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,1536,0.018353778454992507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,128,0.01626044511795044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,10240,1.2309466467963326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,2560,0.011559111376603445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,1024,0.009748444788985783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,12288,0.4509733517964681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,2560,0.020502222908867728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,6144,0.015837333268589444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,7168,0.17335200309753418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,64,0.0033235556135574975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,51200,0.008072000410821702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,3072,0.022318222456508215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,5120,0.4829626613193088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,7168,0.328750212987264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,256,0.003524444583389494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,12288,0.14189778433905706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,128,0.08957599931293064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,6144,0.27549955579969615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,8192,10.388209025065104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,768,0.00870311094654931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,768,0.36043998930189347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,32,0.003428444473279847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,3072,0.05446311169200473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,2560,0.028832889265484277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,3584,0.17947022120157877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,768,0.006979555719428592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,65536,0.26640621821085614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,12288,0.26663467619154185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,128,0.015627556376987033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,51200,0.43504799736870664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,1024,0.012031111452314587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,3584,0.12380354934268528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,768,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,256,0.002959111084540685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,128,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,2560,0.012883555557992725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,10240,0.1844471163219876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,1536,0.05596533086564806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,51200,0.31252800093756783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,2048,0.0210533340771993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,7168,0.009163555171754625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,10240,0.09997422165340847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,8192,0.06260177824232313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,65536,1.054664929707845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,7168,0.5228782229953343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,2560,0.07979644669426812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,1024,0.03354933195643955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,16384,0.23439023229810926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,1536,0.1093920005692376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,3584,0.021173333128293354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,32,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,5120,0.18007555272844103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,8192,0.013430222041077085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,12288,0.019478221734364826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,2560,0.006601777755551868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,128,0.025275554921891954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,128,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,32,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,768,0.0038488888078265717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,3072,0.018017777138286166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,12288,0.06838399834103055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,51200,0.18195111221737334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,2048,2.6490756140814886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,64,0.012690666649076672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,768,0.028903110159768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,3584,0.009304000271691216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,128,0.025204444924990337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,51200,1.8097750345865886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,5120,0.835502200656467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,1536,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,16384,0.168196439743042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,8192,0.04839644498295254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,1024,0.028765334023369685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,32,0.029924445682101782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,10240,0.031601776679356895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,65536,0.2426453431447347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,8192,2.081650627983941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,16384,0.5259920226203071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,32,0.0032915555768542816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,64,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,8192,0.015528000063366361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,256,0.021179555190934077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,512,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,5120,0.25170667966206867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,512,0.006009777800904379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,2048,0.07259022527270846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,8192,0.03011111087269253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,65536,3.040005366007487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,3072,0.035896887381871544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,12288,0.2139271100362142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,3072,0.30280976825290257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,1024,0.004365333252482944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,65536,0.07212711042828031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,128,0.004556444370084339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,8192,0.02681422233581543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,2560,0.05326844586266411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,5120,0.10277955399619208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,64,0.06849422057469685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,1024,0.010433778166770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,4096,0.006343110981914733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,7168,0.010874666273593903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,4096,2.5075431399875217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,64,0.012063110868136087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,16384,1.1231181886461046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,6144,0.013419555293189155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,32,0.005588444570700328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,5120,0.01276533305644989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,6144,0.1807537741131253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,5120,1.4989635679456923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,16384,0.03988622294531928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,2560,0.008072000410821702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,3584,0.01998133295112186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,8192,0.10655288563834296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,65536,0.08548177613152398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,5120,0.011373333632946014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,1024,0.09435644414689805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,5120,0.7318435774909126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,64,0.002540444541308615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,768,0.004866666677925321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,5120,0.04695111182000902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,6144,0.015064888530307345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,1024,0.13779466682010227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,128,0.007232888705200619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,16384,0.2829786671532525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,7168,0.016215999921162922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,10240,0.013644444445768992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,8192,2.008331510755751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,51200,0.12843110826280382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,65536,0.36827201313442653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,1024,0.009317333499590555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,65536,0.09302488962809245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,65536,0.547602653503418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,7168,0.06767111354404025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,8192,0.02126577827665541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,1536,0.01145600030819575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,1536,0.0896400014559428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,5120,0.051474667257732816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,2048,0.018946665856573317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,6144,0.14297155539194742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,512,0.030392888519499037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,128,0.004202666795916027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,64,0.004583999928500918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,65536,11.585484822591146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,6144,0.13887288835313585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,2560,0.028359111812379625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,4096,0.10972177982330322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,65536,0.3547164334191217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,32,0.004716444346639845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,10240,0.02908266584078471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,4096,0.03033422099219428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,128,0.00300799993177255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,64,0.005072888814740711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,2560,0.11366932921939427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,32,0.006989333364698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,12288,0.018704889549149405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,2560,0.00600000015563435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,3584,0.009906666974226633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,51200,0.05997422006395128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,32,0.0025742221623659134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,256,0.0028577778074476454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,1024,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,1536,0.009444444543785518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,3584,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,12288,0.04391111267937554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,1024,0.010644444160991244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,3584,0.011830222275522021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,128,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,3584,0.06438310941060384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,5120,0.02131555477778117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,3584,0.023189332750108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,6144,0.014879110786649915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,3072,0.02390755547417535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,512,0.00499199993080563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,64,0.003137777869900068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,768,0.005681777579916849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,64,0.0034551111360390982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,32,0.004857777721352047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,51200,0.1227102279663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,10240,0.013663999736309052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,2560,0.10101155440012614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,2048,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,512,0.004496888981925117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,1536,1.3475928836398656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,2560,0.010862222148312463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,6144,0.0355057782597012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,256,0.004392000122202767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,3584,0.016789333687888253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,128,0.007232888705200619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,3584,0.024129778146743774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,512,0.0063440000845326324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,5120,0.011609777808189392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,128,0.0035751110149754416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,1536,0.005863111052248213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,256,0.008628444539176093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,2560,0.018292443619834054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,3072,0.007256888680987888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,5120,0.043066667185889355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,5120,0.1865262190500895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,1536,0.0047608887155850725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,1536,0.01861422260602315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,1024,0.1259511046939426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,16384,0.03813777698410882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,10240,0.03302044338650174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,2048,0.009126221968068017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,256,0.00775022225247489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,256,0.0090844440791342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,3072,0.17284354898664686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,12288,0.04415111078156365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,512,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,768,0.0037457777394188773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,5120,0.032834665642844305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,16384,0.01131377783086565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,51200,0.6508471171061198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,51200,3.374592039320204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,3584,0.10210133261150783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,16384,0.0236151119073232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,1024,0.10037866565916274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,16384,0.008294222255547842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,4096,0.00831822223133511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,1024,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,8192,0.39274578624301487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,10240,0.017456889152526855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,5120,0.007620444728268518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,2560,0.007287999822033777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,3072,0.20478667153252494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,5120,0.035629332065582275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,7168,0.2348568969302707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,768,0.005125333451562458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,1536,0.011663999822404651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,16384,0.4052773316701253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,4096,0.011113777756690979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,10240,0.013587555951542325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,768,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,64,0.01923644377125634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,7168,0.081258667839898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,5120,0.018147556318177115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,768,0.022351110974947613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,3072,0.016174222032229107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,10240,0.018441778090265062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,5120,1.4349777433607314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,32,0.006621333460013072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,51200,0.037244445747799344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,64,0.01440622243616316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,6144,1.6894889407687719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,12288,0.24379910363091362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,7168,0.031698667340808444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,8192,0.05394222338994344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,128,0.003012444410059187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,32,0.003342222215400802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,7168,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,6144,0.01426488823360867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,8192,0.031345778041415744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,10240,0.08206222454706828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,768,0.004323555363549126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,4096,0.008814222282833522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,32,0.002628444383541743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,8192,0.030440890126758154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,1536,0.007132444116804335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,3584,0.039150221480263606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,10240,0.05475911166932848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,3584,0.17689689000447592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,768,0.010252444280518426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,5120,0.09856888982984756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,7168,0.16141511334313288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,7168,0.006365333166387346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,1536,2.5249705844455295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,3072,0.011125333607196808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,128,0.016159999701711867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,1024,0.024479111035664875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,1536,0.011063999599880643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,4096,0.012333333492279053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,768,0.005030222237110138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,5120,0.025761778155962627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,128,0.0058711109062035876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,6144,0.013459555804729462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,3072,0.0273324449857076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,3584,0.030476444297366675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,16384,0.1300284465154012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,16384,4.462893168131511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,32,0.018587556150224473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,65536,0.1878960000144111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,3072,0.018033777674039204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,6144,0.07165955834918551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,10240,0.526077323489719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,6144,0.037351999017927386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,2560,0.26473244031270343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,256,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,64,0.006688000013430913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,12288,0.017701332767804463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,4096,0.22618044747246635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,65536,0.6487911012437608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,8192,0.21483200126224092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,4096,0.009945777555306753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,32,0.005981333139869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,3072,0.25960977872212726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,7168,0.023386667172114056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,8192,0.0224897778696484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,8192,0.30902311537000865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,16384,0.04661866691377428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,51200,0.46761157777574325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,6144,3.0595253838433156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,12288,0.04252444373236763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,256,0.011599111060301462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,51200,1.023528840806749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,8192,0.15557155344221327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,12288,0.0203840004073249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,16384,0.035755554835001625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,51200,0.1112826665242513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,6144,3.3888337877061634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,7168,0.07048355870776706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,64,0.0033191111352708605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,8192,0.019465777609083388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,3584,0.036426666710111834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,128,0.01073688848151101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,1024,0.02199466692076789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,3584,0.008833777573373582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,6144,2.5229600270589194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,51200,0.07917244566811456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,7168,0.013266666895813413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,32,0.008184889124499427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,16384,0.0328764451874627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,5120,0.27034399244520396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,256,0.0031315556002987754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,6144,0.027860442797342937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,51200,0.11733155780368382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,12288,0.09607911109924316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,51200,0.20134488741556802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,32,0.002797333316670524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,64,0.011810666984981961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,65536,3.564239078097873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,16384,0.050586667325761586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,6144,0.11165600352817112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,2560,0.0277173337009218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,1024,0.08991911013921101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,4096,0.0798106657134162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,7168,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,12288,0.009336888790130615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,1536,0.05176622337765164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,12288,0.018072888255119324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,8192,0.061576002173953585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,128,0.03031022349993388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,16384,2.6396329667833114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,128,0.003955555458863576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,3584,0.021811554829279583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,768,0.27958843443128795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,32,0.03510933452182346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,7168,3.337974124484592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,1536,0.007989333735571967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,1024,0.07102311319775052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,256,0.005165333135260476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,32,0.041137778096728854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,16384,0.03185155656602647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,128,0.006721777634488211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,32,0.0028008888992998335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,768,0.2669911119672987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,5120,0.04528888728883532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,10240,0.01514755520555708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,3584,0.011076444553004371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,51200,1.7721715503268773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,3584,0.009293333523803288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,1536,0.03275555703375075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,10240,0.03666222095489502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,512,0.005492444253630108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,768,0.026332444614834253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,32,0.003546666767862108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,1536,0.016900445024172466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,12288,0.012787555654843649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,512,0.005664000080691443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,1536,0.1466577715343899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,8192,0.30191556612650555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,2560,0.041866666740841336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,6144,0.15653333399030897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,128,0.004794666750563515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,2560,0.18915466467539468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,51200,0.7476035753885905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,16384,0.17394310898251006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,6144,0.1325182252460056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,1536,0.004670222186379963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,5120,0.03517511155870225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,1536,0.07990133100085788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,128,0.004859555512666702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,4096,0.028739555014504328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,256,0.015473778049151102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,8192,0.1590826643837823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,5120,0.01015644437736935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,65536,0.19931199815538195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,8192,0.12037332852681477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,10240,0.048546665244632296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,1024,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,256,0.02935289012061225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,4096,0.016270221935378183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,2048,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,3072,0.00833155545923445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,256,0.01943288909064399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,2560,0.024490666058328416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,3584,0.007932444413503012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,256,0.015840000576443143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,4096,0.008210666477680206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,3584,0.007990222010347579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,64,0.002879111096262932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,2048,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,1024,0.012682666381200155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,256,0.02055111030737559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,4096,0.2515457736121284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,256,0.0092284447617001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,10240,0.015875554747051664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,12288,0.31781334347195095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,12288,0.01587199999226464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,2048,0.01050222251150343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,2048,0.10702755716111924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,65536,1.4143760469224718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,2560,0.013224889006879596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,32,0.006455111006895701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,6144,0.16076710489061144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,1024,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,128,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,768,0.0036480000449551474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,768,0.028871999846564397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,1024,0.006106666806671355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,1024,0.003699555579158995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,51200,1.4211919572618272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,51200,17.6572265625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,12288,0.1310675541559855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,4096,0.03133066826396518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,128,0.0064479998416370815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,4096,0.706615130106608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,8192,0.18322577741410997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,12288,0.08215910858578153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,65536,0.20652088854047987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,10240,0.03340355555216471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,8192,0.21612800492180717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,128,0.08717244201236302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,128,0.0034755555291970572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,5120,0.039727999104393855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,12288,0.02444000045458476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,5120,0.022203556365436975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,4096,0.02470933397610982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,768,0.009358222285906473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,256,0.0034497777620951333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,8192,0.03872977693875631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,12288,0.03798933492766486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,128,0.01130844404300054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,1024,0.0058648888435628675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,16384,0.8022266493903266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,8192,0.9300595389472114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,8192,0.07288889090220134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,3584,0.016521778371598985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,768,0.00868977771864997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,51200,0.4587253464592828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,1536,0.030071109533309937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,1024,0.006184000107977126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,2560,0.02457777824666765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,4096,0.009824000298976898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,7168,0.04762399858898587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,16384,0.04051911168628269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,5120,0.1914773384730021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,512,0.005504000104135937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,65536,0.19753689236111113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,1536,0.10385333167182074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,64,0.0027822222974565295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,8192,0.21307378345065645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,6144,0.006073777874310811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,256,0.005746666755941179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,1536,0.1263048913743761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,65536,0.23499290148417154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,256,0.005748444547255834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,128,0.003334222154484855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,51200,0.26854488584730357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,5120,0.00868977771864997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,10240,0.21308978398640951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,256,0.00407644444041782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,51200,0.580261336432563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,16384,0.36420178413391113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,12288,0.018516444497638278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,64,0.0028897778441508612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,12288,0.021000888612535264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,1536,0.010808000134097205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,3584,0.25909601317511666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,65536,0.1293333371480306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,256,0.1521351072523329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,64,0.002874666617976295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,16384,0.5612159834967719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,12288,0.24182576603359648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,3072,0.015174221661355762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,4096,0.01872177753183577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,2560,0.05136533247099983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,768,0.00942666663063897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,768,0.05983555316925049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,2048,0.009201777478059133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,3072,0.012089778151777057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,10240,0.03848000036345588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,2560,0.07334578037261963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,512,0.028247998820410833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,16384,1.3007449044121635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,1024,0.007629333270920648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,1024,0.011724444727102915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,7168,0.2409902148776584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,5120,0.010807111031479306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,12288,0.019704888264338177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,512,0.019677332705921598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,512,0.00902844468752543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,10240,0.033048888047536217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,65536,0.12503555085923937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,1024,0.009756444229020013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,3072,0.007975999679830339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,3072,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,4096,0.0159244438012441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,32,0.025215999947653875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,1024,0.06258400281270345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,8192,0.03657866517702738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,64,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,32,0.005191111316283544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,51200,5.7029207017686625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,3584,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,51200,2.634765413072374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,2560,0.05644355879889595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,1024,0.022630222969584998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,1024,0.021035555336210463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,4096,0.2006604406568739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,2048,0.007890666524569193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,768,0.04231466518508064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,512,0.01900533338387807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,768,0.010240000155236986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,12288,0.08719822433259751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,2048,0.009343110852771336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,65536,13.814764234754774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,64,0.013291555146376291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,65536,0.0769324435128106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,1024,0.03266755408710904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,65536,1.6509955724080403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,256,0.01166044423977534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,2560,0.47058577007717556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,4096,0.07523822122149997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,768,0.009480000370078618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,2560,0.01850044396188524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,32,0.007920000288221572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,1536,0.007642666498819987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,65536,0.5730755594041612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,6144,0.024751999311976965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,3072,0.018331555856598746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,16384,0.12400978141360813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,3584,0.01126933346192042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,512,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,2048,0.04728977878888448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,256,0.15405155552758112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,32,0.034544888469907976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,512,0.003779555360476176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,8192,0.018894222047593858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,128,0.002740444408522712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,10240,0.0789324442545573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,3584,0.006659555352396435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,6144,0.16898132695092094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,5120,0.7538142204284668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,1024,0.009768000079525841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,128,0.002805333377586471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,16384,0.036179555786980525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,7168,0.011053333679835001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,12288,0.18489155504438612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,51200,1.5586018032497828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,51200,0.22027822335561117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,1024,0.014476444986131458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,10240,0.07662044631110297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,8192,0.01975555552376641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,3072,0.9587973488701714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,5120,0.011352000137170156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,2560,0.03848355677392747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,10240,0.4417591094970703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,8192,0.017052445146772597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,3584,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,5120,0.009962666365835402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,2048,0.006631999793979857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,3072,0.0869999991522895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,16384,1.6447324752807617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,16384,0.02511466708448198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,4096,0.012466666599114736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,65536,107.48567030164931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,32,0.1906977759467231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,2560,0.00795377790927887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,2048,0.030738668309317693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,768,0.16635466946495905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,1536,0.00656355544924736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,7168,0.23684623506334093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,256,0.004102222207519743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,10240,0.00941777808798684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,3072,0.019205333458052743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,51200,0.6593342357211643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,2560,0.01737333337465922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,8192,0.15028356181250677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,6144,0.0413315561082628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,128,0.00851733320289188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,8192,0.040592888991038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,16384,0.16220445103115505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,512,0.003469333259595765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,3072,0.043121778302722506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,2560,0.028888887829250757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,32,0.016545777519543965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,128,0.0028355556229750314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,1536,0.013908444179428948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,3072,0.015253333581818474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,2048,0.010197333163685268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,65536,6.040999094645183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,10240,0.050960001018312245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,256,0.006647111227114995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,16384,0.0351182222366333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,64,0.006191111273235745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,512,0.01566755606068505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,2048,0.17735111713409424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,32,0.20037689473893908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,512,0.0035084444615576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,1024,0.011391999820868174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,1536,0.10610222154193455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,16384,0.1335759957631429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,65536,0.4090542263454861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,1024,0.010392888552612728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,1536,0.0148044443792767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,16384,0.7136809031168619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,7168,0.8818702167934842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,7168,0.0166284441947937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,4096,0.01072533345884747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,3584,0.026720888084835474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,256,0.004753777964247597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,128,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,1536,0.017119111286269296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,2048,0.14197777377234563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,51200,0.06522133615281847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,6144,0.013112000293201871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,7168,0.2288142310248481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,6144,0.0391226675775316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,65536,0.06890310843785603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,32,0.015421332584487068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,12288,0.2523599995507134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,16384,0.29484089215596515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,1024,0.016866667403115165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,10240,37.042134602864586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,768,0.17501955562167695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,1536,0.07578311363855998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,4096,0.2830924457973904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,12288,0.04817777872085571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,16384,0.05551022291183472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,768,0.006062222023804982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,4096,0.0598542226685418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,6144,0.07271555397245619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,256,0.003006222140457895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,1024,0.02453511125511593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,16384,0.26393244001600474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,16384,0.027649778458807204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,10240,0.009147555463843876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,2048,0.009780444204807281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,64,0.009463999834325578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,128,0.01495733360449473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,4096,0.009668444593747457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,12288,0.05357066790262858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,512,0.023834667272037927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,3072,0.011819555527634092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,16384,0.04872177706824409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,12288,0.04279466801219516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,12288,0.47555287679036456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,512,0.006291555447710886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,4096,0.011129777464601727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,768,0.0042151109211974675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,12288,0.04144355654716492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,7168,0.014849777022997538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,32,0.0028231110837724474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,3072,0.022421333524915908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,65536,0.06057155794567532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,51200,1.6687244839138453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,32,0.0028400001012616684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,256,0.007458666960398356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,12288,0.2586328983306885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,768,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,512,0.003593777616818746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,768,0.008866666919655269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,10240,0.18334578143225777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,2560,0.024375110864639282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,1536,0.01012711144155926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,3584,0.012312889099121094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,2048,0.5360150867038304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,768,0.010379555324713389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,3584,1.271351072523329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,8192,0.3662719991472032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,3584,0.15381156073676214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,128,0.018429333964983623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,65536,0.08341866731643677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,3072,0.009579555855857002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,1536,0.022403554783927068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,512,0.005495999836259418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,10240,0.4660968780517578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,2560,0.06477599673800997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,3584,0.07642844650480482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,32,0.0028497777465316984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,2560,0.018392889036072623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,2560,0.024497777223587036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,65536,0.09533955653508504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,256,0.399999115202162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,64,0.003589333345492681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,32,0.0026364444444576898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,512,0.005233777893914117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,7168,0.11544088522593181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,1536,0.008016889293988546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,128,0.010379555324713389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,2048,0.005912888795137405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,5120,1.002571529812283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,2048,9.460161844889322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,1536,0.004679999831649992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,3584,0.09317511320114136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,4096,0.016711110870043438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,16384,0.1255813307232327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,7168,0.0403439998626709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,3584,0.022279999322361414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,768,0.005691555639108022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,128,0.017089777522616915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,4096,0.026475555366939966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,7168,0.018466666340827942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,3072,0.016819554898473952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,64,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,5120,0.016520889268981088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,1024,0.14596088727315268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,64,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,256,0.0046302220887608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,12288,0.006128888991143968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,128,0.6443831125895182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,5120,0.015210666590266757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,6144,0.012467555701732635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,3072,0.08916799889670478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,64,0.04268355502022637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,6144,0.18060621950361463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,3584,0.10622044404347737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,5120,0.25028978453742134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,512,0.04055733482042948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,2560,0.011582222249772815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,10240,0.09159911341137356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,32,0.003040888864133093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,2560,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,768,0.023294222023752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,2048,0.006762666834725275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,3072,0.00997688869635264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,512,0.007090667055712805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,65536,0.06732889016469319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,7168,0.008258666429254744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,10240,0.048264887597825795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,3072,0.011404444773991903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,51200,0.3252577781677246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,51200,0.20626844300164116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,64,0.005120000077618493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,32,0.00291377781993813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,12288,0.411191119088067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,65536,3.5254569583468967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,2048,0.17177155282762316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,2560,0.00793511089351442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,65536,0.1077422234747145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,1024,0.011425777441925473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,8192,0.8201200167338053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,16384,0.07856800158818562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,2560,0.017441777719391715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,2048,0.008472888833946651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,128,0.014941333068741692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,16384,0.02472977836926778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,5120,0.08772266573376125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,3584,0.34856266445583767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,64,0.002983999955985281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,2048,0.046441776884926685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,12288,5.5544221666124125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,51200,0.11346044805314805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,768,0.014003554979960123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,1536,0.01794488893614875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,3072,0.008219555848174626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,10240,0.14759022659725612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,2048,0.009686222506894005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,256,0.004845333182149463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,5120,0.06993244753943549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,1536,0.006491555521885554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,51200,0.1787431107627021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,3072,0.009891555541091496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,256,0.005500444521506627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,1536,1.5681315528021917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,256,0.016179554992251925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,16384,1.2381582260131836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,64,0.0096124443742964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,512,0.01036800030204985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,256,0.00692622239391009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,1024,0.017652443713612027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,768,0.007707555260923173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,3072,0.03234666585922241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,51200,0.19611288441552055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,3584,0.42976978090074325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,6144,0.07608711057239108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,1536,0.01680444512102339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,65536,0.01757688820362091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,512,0.028948442803488836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,768,0.037760890192455716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,512,0.09171644184324478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,32,0.0030088888274298776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,2560,0.011485333244005838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,2560,1.0553937488132052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,51200,0.11229956150054932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,65536,0.07313422362009685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,3072,0.013743999931547375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,1024,0.006871999965773688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,512,0.006098666538794835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,256,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,12288,0.03821422325240241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,3584,0.021456888980335657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,10240,0.045775112178590566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,128,0.017271111408869427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,2560,0.09245689047707452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,64,0.032060444355010986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,32,0.002576888849337896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,128,0.022875555687480505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,1024,0.015595555305480957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,32,0.008824889030721452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,8192,0.39211376508076984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,3072,0.007191110816266801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,2560,0.013334222137928009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,64,0.009030222064918941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,7168,0.08042666647169325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,3072,0.008339555727110969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,65536,15.617253621419271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,65536,0.3360995451609294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,6144,0.030252443419562444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,5120,0.02420622275935279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,65536,0.07552622424231635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,5120,0.382421334584554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,6144,0.11588888698154026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,2560,0.0906640026304457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,32,0.18951643837822807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,10240,0.010486221975750394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,3584,0.06856622298558553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,12288,0.05532977647251553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,1024,0.0057955553962124725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,2560,0.00997155573632982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,5120,0.09857155879338582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,1536,0.15871555275387234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,6144,0.01455200049612257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,16384,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,256,0.002974222310715251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,128,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,3584,0.0138568886452251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,1024,0.003946666502290302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,512,0.03977333174811469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,16384,0.019356444478034973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,2048,0.03728355632887946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,5120,0.008919999831252629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,65536,0.4323626624213324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,2048,0.21388088332282174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,12288,0.028034667174021404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,1024,0.23691021071539986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,2560,0.00703288863102595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,2560,0.023038221730126276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,51200,0.3300088776482476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,3072,0.08248888783984713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,512,0.01633155511485206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,512,0.02513066596455044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,256,0.06705866919623481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,2560,0.010130667024188572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,1024,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,3072,0.009174221919642555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,128,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,5120,0.009771555662155151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,512,0.00489777781897121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,8192,0.01776088939772712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,64,0.005384888913896348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,2048,0.01126488877667321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,5120,0.009491555392742157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,128,0.0035706667436493766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,2560,0.20761688550313315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,768,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,5120,0.011243555280897351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,3072,0.017179555363125272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,32,0.008467555873923832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,768,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,3584,0.038183109627829656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,10240,0.29941778712802464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,3072,0.31828710767957896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,3072,0.013171555267439948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,8192,0.020271110865804885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,1536,0.04053599966896905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,256,0.01089599976936976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,768,0.006545777950021956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,10240,0.04654133319854736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,5120,0.03378488951259189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,10240,1.5795893139309352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,512,0.008417777717113495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,64,0.04912177721659342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,32,0.02330222229162852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,4096,0.02348888913790385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,10240,0.0075253330998950535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,4096,0.030563556485705908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,1024,0.006864000111818314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,8192,0.014164444473054675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,2048,0.028276443481445312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,256,0.008964444200197855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,512,0.011180444724029966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,6144,0.045257776975631714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,51200,17.306532118055554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,12288,0.019985778464211356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,128,0.10933599869410197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,3072,0.06756711006164551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,12288,0.03426222337616815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,256,0.004579555657174853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,3584,0.08424533075756496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,2048,0.15898222393459743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,8192,0.025936888323889837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,1536,0.6392791006300185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,512,0.0046924443708525766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,5120,0.04462755719820658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,1024,0.038944890101750694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,12288,1.1313759485880535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,64,0.005482666608360078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,6144,0.03517599900563558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,12288,8.885409037272135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,1536,0.00909511082702213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,51200,0.04962311188379923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,51200,2.8079954783121743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,128,0.016571554872724745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,7168,0.024102222588327196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,16384,0.41800443331400555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,3584,0.12200178040398492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,51200,0.17941955725351968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,64,0.0068195557428730865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,7168,0.7462435298495822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,768,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,256,0.0046880000995265115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,4096,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,51200,0.058196445306142174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,256,0.011053333679835001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,51200,0.6021484269036187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,1536,0.009729777773221334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,2560,0.007962666451931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,256,0.005233777893914117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,128,0.002921777880854077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,256,0.017473777135213215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,128,0.002951111023624738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,12288,0.029398222764333088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,3584,0.047094222572114736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,16384,0.1257751120461358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,3584,0.01819022165404426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,128,0.005351999981535806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,2560,0.012345777617560493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,16384,4.7859475877549915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,51200,0.42960532506306964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,12288,0.06288444333606295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,512,0.01202755586968528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,65536,0.9883955849541558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,12288,0.2434746689266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,1536,0.027401778433057997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,256,0.011500444677140979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,65536,0.08053421974182129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,1536,0.05983289082845052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,8192,0.0191600008143319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,256,0.008390222158696916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,12288,0.08550577693515354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,2048,0.01591466698381636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,12288,0.01978488928741879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,16384,0.39914578861660427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,10240,0.04216177927123176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,5120,0.008147555920812819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,768,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,10240,0.1570391125149197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,12288,0.022934221559100684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,768,0.033511112133661904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,10240,0.015749333633316886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,64,0.01092088884777493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,12288,0.03537155522240533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,12288,0.02882222334543864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,7168,0.057461334599388965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,10240,0.01348533315791024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,3584,0.09306044711007012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,51200,0.7253360218471951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,6144,0.01861422260602315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,768,0.005509333478079901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,2560,0.4422977765401204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,10240,0.016930666234758165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,512,0.0035075553589397003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,12288,0.13731911447313097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,3072,0.08816355466842651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,10240,1.2328159544203017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,4096,0.020045333438449435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,10240,0.3402577771080865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,256,0.008995555341243744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,8192,0.03472444415092468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,4096,0.007297777467303806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,3072,0.011593778100278644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,1024,0.11727200614081489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,2560,0.03086755673090617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,64,0.012147555748621622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,256,0.017441777719391715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,512,0.01793688866827223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,256,0.002982222164670626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,7168,0.019766221443812054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,2560,0.01905955539809333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,2560,0.10032977660497029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,3584,0.059487998485565186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,1536,0.008065777520338694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,16384,0.023500444160567388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,2048,0.12179111109839545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,5120,0.010457778142558204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,8192,1.155443509419759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,2560,0.04380888740221659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,768,0.0060159998635451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,64,0.02576711111598545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,6144,0.014388443695174323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,4096,0.024805333879258897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,128,0.008558221989207799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,12288,0.5919200049506294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,51200,67.54224989149306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,8192,1.6567742029825847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,256,0.027678221464157104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,16384,0.03661422265900506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,32,0.007663999994595845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,3584,0.3371688789791531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,10240,0.0221760008070204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,3072,0.008340444829728868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,768,0.010715555813577441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,128,0.18519732687208387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,3072,0.015674667225943673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,256,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,3584,0.015085332923465304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,128,0.005240889059172736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,512,0.00811822215716044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,3072,0.1893155574798584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,10240,0.02646222213904063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,10240,1.037769741482205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,16384,0.09503999683592056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,3072,0.0831422209739685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,7168,0.28339378039042157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,512,0.005836444596449534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,51200,0.0634960002369351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,10240,2.056884341769748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,8192,0.016014221641752455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,7168,0.016872889465755887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,32,0.005953777581453323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,7168,0.014712888333532544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,16384,0.4758355352613661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,512,0.06345333655675252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,3584,0.014109333356221518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,32,0.00776622196038564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,3072,0.08080266581641303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,10240,0.8586293326483833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,16384,0.13566133711073133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,8192,0.05230844351980421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,256,0.13986755741967097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,1024,0.007879111501905654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,3584,2.42777591281467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,3584,0.06057066387600369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,5120,0.046959112087885536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,2048,0.00812533332241906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,12288,0.04166933231883579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,16384,0.13570933871799046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,51200,0.05451733205053541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,51200,0.4192684491475423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,12288,0.057723555299970836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,3584,1.4880524741278753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,2560,0.01978933314482371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,1536,0.02529866662290361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,128,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,1024,0.007499555746714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,512,0.00868711123863856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,16384,0.08289155695173475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,128,0.13747466935051814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,12288,0.2550506591796875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,768,0.004605333424276776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,1536,0.015616888801256815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,7168,0.35434667269388836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,3584,0.13277244567871094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,64,0.003120888852410846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,1024,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,51200,13.960601806640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,512,0.023630221684773762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,12288,0.02792000108295017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,128,0.007058666812049017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,7168,0.02079288827048408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,4096,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,256,0.005610666755172942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,3072,0.008548444343937768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,2048,0.09089689122305976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,256,0.00556355549229516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,5120,0.26476976606580943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,128,0.016915554801623028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,6144,0.14755911297268337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,16384,0.022894221875402663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,256,0.003060444361633725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,1024,0.021224000387721594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,12288,0.017460443907313876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,2048,0.06415644619199964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,1536,0.005879999862776862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,6144,0.1384506622950236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,7168,0.090756442811754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,2048,0.03591377867592706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,5120,0.44207199414571124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,51200,0.9749431610107422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,2560,0.010568000376224518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,32,0.010657777388890585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,2560,0.030692444907294378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,512,0.01459022197458479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,2048,0.8454897668626574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,65536,0.2852684391869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,12288,0.023974221613672044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,12288,0.037765334049860634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,10240,0.11628444989522298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,16384,0.05501688851250542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,65536,0.8728204303317599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,1536,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,6144,0.03461066550678677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,3584,0.10516711076100667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,4096,0.026593777868482802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,128,0.0058017778727743365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,768,0.00500177757607566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,1536,0.18392444981469047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,1024,0.012989333106411828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,128,0.007551111280918121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,2560,0.18698133362664116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,2048,0.008477333519193862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,256,0.0036355555057525635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,2560,0.012166221936543783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,10240,0.015397333436542086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,4096,0.014779556128713818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,2560,0.009485333330101436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,6144,0.18852621979183623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,768,0.005722666780153911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,768,0.00583022211988767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,7168,0.8947920269436307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,12288,15.522369384765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,1536,0.026143999563323125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,2048,0.47336800893147785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,2048,0.0074622225430276655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,1536,0.015885333220163982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,65536,0.0973893337779575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,256,0.0028435554769304064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,16384,0.07228266530566745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,51200,0.0581608878241645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,2560,0.030018667380015056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,768,0.02353333267900679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,1024,0.041971554358800255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,768,0.00378933341966735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,5120,1.7085644404093425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,1024,0.22488265567355684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,3584,0.00944088896115621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,2048,0.006385777973466449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,16384,83.76358371310764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,768,0.007703999678293864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,512,0.0656871133380466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,2560,0.017343110508388944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,16384,0.27883556154039174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,2560,0.009845332966910468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,768,0.028079999817742243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,768,0.019754666421148512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,7168,0.013911111487282647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,3584,0.010560000108347999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,1024,0.0063546668324205614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,6144,0.1965351104736328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,2560,0.26311198870340985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,12288,0.07935288879606459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,7168,0.02367288867632548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,768,0.008907555705971187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,768,0.01860000027550591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,7168,0.4441804356045193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,64,0.015690666106012132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,7168,0.0255031105544832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,51200,0.4422453244527181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,10240,0.02314044369591607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,16384,3.8256746927897134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,768,0.00982488857375251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,4096,0.15068266126844618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,128,0.005571555760171678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,10240,0.0069226668112807805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,2560,0.020267556111017864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,3584,0.010673777924643623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,5120,2.09208615620931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,12288,0.008784000244405534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,51200,0.041858666472964816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,8192,0.5584915479024252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,256,0.004300444490379757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,8192,0.018208889497650992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,10240,0.01516888870133294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,256,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,128,0.0046471113132105935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,64,0.007205333146784041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,3584,0.13221155272589788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,1024,0.006672888994216919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,7168,0.1326026651594374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,64,0.0027635554886526535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,8192,0.010314666562610203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,256,0.006497777584526274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,5120,0.13760888576507568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,51200,0.14712266127268472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,16384,3.313581254747179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,768,0.007992000215583378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,1024,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,768,0.00572622236278322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,3072,0.01019022199842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,2048,0.007637333538797166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,8192,0.014907555447684394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,256,0.005997333261701796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,3072,0.00738933351304796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,51200,0.4152533213297526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,8192,0.011703111231327057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,1024,0.00416355538699362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,6144,0.7981280220879449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,5120,0.03832888934347365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,7168,0.025993777645958796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,2048,0.011736888852384357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,8192,0.035720000664393105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,10240,3.0244079165988498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,512,0.01893422173129188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,65536,0.3013857735527886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,6144,0.16051822238498265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,16384,0.20746844344668916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,51200,0.6414533191257054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,256,0.012246222131782107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,6144,0.03576622075504727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,65536,4.870894114176433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,65536,0.1592737833658854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,768,0.053563555081685386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,16384,0.06726844443215264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,4096,0.09771555662155151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,1024,0.039515554904937744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,7168,0.235917329788208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,32,0.005585777676767773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,6144,0.2604657808939616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,512,0.0037831109431054858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,256,0.008194666769769456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,7168,0.633344915178087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,256,0.006959111326270633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,768,0.005069333232111401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,4096,0.012601777911186218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,10240,0.2609066698286268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,51200,0.2507351239522298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,64,0.014696000350846184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,5120,1.6568080054389105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,128,0.00812533332241906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,7168,0.23155999183654785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,65536,0.0688471131854587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,16384,0.06836622291141085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,3072,0.016368000043763053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,3584,0.009359999663299983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,512,0.00555288874440723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,32,0.026205332742797002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,1024,0.2779128816392687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,51200,0.07613777452045016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,1024,0.00397777764333619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,3072,0.010654222634103564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,256,0.00309955556359556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,12288,0.07777422004275851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,1536,0.0052266667286554975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,16384,0.011726222104496427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,51200,0.06295555830001831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,32,0.010799999866220685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,2048,0.09495911333296035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,3584,0.025682666235499915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,512,0.004830222162935468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,10240,0.08399822314580281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,8192,0.03126044405831231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,16384,0.026927111877335444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,512,0.014880888991885714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,128,0.009467555416954888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,1024,0.012190221912331052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,128,0.006183111005359226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,2560,0.01868533260292477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,12288,0.0306275553173489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,128,0.0058293334311909145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,65536,2.4697581397162542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,1536,0.007848000360859765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,1024,0.011791110866599612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,1536,0.007520000139872233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,1536,0.23307556576199004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,2560,0.11371733082665338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,1536,0.03275288807021247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,16384,0.14424355824788412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,10240,0.008310221963458592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,2560,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,7168,0.07183288865619235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,1024,0.016199999385409884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,5120,0.021853332718213398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,2048,0.012183999849690331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,7168,0.10564977592892116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,1536,0.005506666584147348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,51200,0.06959555546442668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,6144,0.6667093171013726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,51200,0.18955466482374403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,2048,0.007110222346252865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,10240,0.00979200005531311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,2048,0.012152888708644442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,6144,0.02018044392267863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,5120,0.012454222473833295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,64,0.01295644458797243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,768,0.07860800292756823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,2048,0.04024444354905023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,1024,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,16384,0.006068444500366847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,51200,0.06792977783415052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,2048,0.020136000381575692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,6144,0.03329688972897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,10240,0.8462382422553169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,256,0.0036071112586392295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,8192,0.1305333375930786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,128,0.0028000000036425064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,10240,0.01644177734851837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,4096,0.01018133345577452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,65536,4.3550292121039496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,8192,0.023572444915771484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,12288,0.03415022293726603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,512,0.009600888523790572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,12288,0.09471022420459324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,6144,0.05355022350947062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,2048,0.2657333215077718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,12288,0.018734221657117207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,256,0.04038844505945841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,1024,0.00741244438621733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,32,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,7168,0.03028977910677592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,7168,0.03880177934964498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,256,0.03547733359866672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,7168,0.40563734372456867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,5120,0.5719208717346191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,51200,0.12007910675472683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,6144,0.13748355706532797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,1024,0.027595554788907368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,1024,0.13219288984934488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,2560,0.022237333986494277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,64,0.003387555480003357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,8192,0.1928559939066569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,2048,0.008385777473449707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,10240,1.097992049323188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,64,0.008800889054934183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,1536,0.008736888567606607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,128,0.1854471100701226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,65536,0.3715928925408258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,8192,0.045680890480677284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,512,0.008375999828179678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,51200,2.9188736809624563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,128,0.0032791110376516977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,5120,0.01940000057220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,7168,0.505684428744846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,4096,0.017364444004164804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,5120,0.0123511114054256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,7168,0.020534222324689228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,256,0.1394124428431193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,8192,0.03337422344419692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,3072,0.009215110705958473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,8192,0.1544871065351698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,51200,0.48377423816257054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,1536,0.04547911220126682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,32,0.00370399985048506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,512,0.008529778156015608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,65536,5.118935902913411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,10240,0.05776266919242012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,3584,0.048863109615114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,16384,0.036023110151290894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,7168,0.24487821261088052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,7168,0.01918044520748986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,32,0.0052915554907586836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,65536,6.551652272542317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,10240,0.029840889904234145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,10240,0.05342400074005127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,3072,0.006550222221348021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,256,0.00554577757914861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,32,0.0036151111125946045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,128,0.003914666672547658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,128,0.006694222076071634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,8192,0.2680186695522732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,2048,0.11051199833552043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,51200,2.6323174370659723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,6144,1.2539289262559679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,16384,0.05526577764087253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,12288,0.0317831105656094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,3584,0.10379733641942342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,7168,0.05264000097910563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,2560,0.09128355317645603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,256,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,6144,0.014056000444624158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,2048,0.011095110740926532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,2048,0.006782222125265334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,65536,0.2015697823630439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,32,0.12603021992577448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,256,0.002842666581273079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,128,0.008554666406578487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,65536,4.865010579427083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,7168,1.0923404693603516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,8192,0.018053332964579265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,8192,0.017940445078743827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,7168,0.409988456302219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,5120,0.21588888433244494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,65536,1.9113911522759333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,32,0.003636444608370463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,16384,0.034252444903055825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,10240,0.04764622118737963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,16384,0.0705839991569519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,6144,0.01851733360025618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,12288,0.07589955462349786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,32,0.01575911045074463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,65536,4.006578657362196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,3584,0.020407110452651978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,256,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,5120,0.010431110858917236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,12288,0.12909777959187826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,512,0.018830221560266282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,51200,0.0602453351020813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,2048,0.007816000117195977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,32,0.0057653333577844836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,4096,0.015636444091796875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,65536,0.24842754999796549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,1024,0.175653338432312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,2048,0.007073777417341868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,6144,0.055008888244628906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,1536,0.12586577733357748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,3584,0.01789955629242791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,16384,0.07512178023656209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,10240,0.07464444637298584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,6144,0.030240890052583482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,32,0.027143999934196472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,3072,0.01626133256488376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,1024,0.03388533327314589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,768,0.00565422202150027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,2560,0.042619556188583374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,1536,0.019734222027990553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,1024,0.034210665358437434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,256,0.003043555551105075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,12288,0.012680889003806643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,8192,0.025376889440748427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,51200,0.39384622044033474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,512,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,4096,0.10225066873762344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,8192,0.03754577702946133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,8192,0.26931556065877277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,65536,0.17787022060818145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,1536,0.004483555754025777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,1024,0.005443555613358815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,3072,0.01368977791733212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,1536,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,32,0.020994666549894545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,8192,0.021143111917707656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,128,0.014906666345066495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,6144,0.03029511041111416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,6144,0.1183591153886583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,256,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,128,0.046253333489100136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,2560,0.01478844384352366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,2048,0.008218666745556725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,5120,0.10288711388905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,1024,0.02777422302299076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,4096,0.1404204501046075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,8192,0.03903911179966397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,65536,0.06633422109815809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,5120,0.09585866663191055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,12288,0.16536000039842394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,3584,0.17211911413404676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,1024,0.008015111088752747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,7168,0.014914666612943014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,12288,0.13020888964335123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,16384,0.061160888936784535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,51200,0.07305866479873657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,4096,0.010077333284748925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,256,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,7168,0.022832888695928786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,1536,0.00848711116446389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,2048,0.030819555123647053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,7168,0.053732444842656456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,10240,0.018567999203999836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,5120,0.023660444551044043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,3584,0.026679111851586237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,32,0.00508266646001074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,51200,0.2876906659868029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,512,0.013235555754767524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,1024,0.04081333345837063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,128,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,6144,0.022103110949198406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,4096,0.012229333321253458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,256,0.004620444443490771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,8192,0.014949333336618213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,2560,0.007448888487286038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,32,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,7168,0.009161777794361115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,6144,0.1135973268085056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,10240,0.03967466619279649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,128,0.009365333451165093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,3072,0.01124888906876246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,16384,0.021764443980322942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,4096,0.006780444333950679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,2560,0.005980444451173146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,5120,0.1412960025999281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,2048,0.5799999766879612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,3584,0.01204711116022534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,5120,0.03872888949182298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,65536,0.40773865911695695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,256,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,7168,0.021528888079855178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,3584,0.019367110398080613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,32,0.03047555685043335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,4096,0.018555555078718398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,51200,0.12887110975053576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,7168,0.03412088751792908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,5120,0.24226665496826172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,1024,0.025949334104855854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,2048,2.1272303263346353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,7168,0.013859555953078799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,12288,0.33189333809746635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,4096,0.021457778082953557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,16384,0.9658275180392795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,5120,0.015319999721315173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,1536,0.04385777645640903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,8192,0.04828977915975782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,768,0.009983110758993361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,7168,0.49202844831678605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,65536,0.493327988518609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,128,0.006222222414281633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,1536,0.037608888414171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,768,0.008606221940782335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,2560,0.019088000059127808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,7168,0.02416977783044179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,51200,0.13329244984520808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,51200,20.63016594780816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,512,0.005568888866239124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,2560,0.00832266691658232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,768,0.1110888918240865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,3072,0.051552888419893056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,2048,0.012163555456532372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,1024,0.031340443425708346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,3072,0.007808888951937358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,7168,0.02023555503951179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,5120,0.021496888664033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,4096,0.08702311250898574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,1024,0.010849778023031024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,65536,0.1413431167602539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,256,1.146808836195204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,768,0.00943999985853831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,7168,0.030061334371566772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,3584,0.014727110664049784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,2048,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,7168,0.14960355228847927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,768,0.021035555336210463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,10240,0.05363377928733826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,8192,0.024136889312002394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,768,0.037046220567491316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,16384,0.034105777740478516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,5120,0.01036444471942054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,65536,0.06617599725723267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,65536,0.5308062235514323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,2560,0.012182222472296821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,65536,1.591933356391059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,4096,0.02145333256986406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,8192,0.248705784479777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,7168,0.4279342227511936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,1536,0.019288000133302476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,4096,0.025792888469166223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,2048,0.007866666548781926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,4096,0.005975111077229182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,768,0.1269715494579739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,512,0.009714666340086196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,12288,0.018836443622907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,768,0.01038933296998342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,2048,0.007135999699433644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,768,0.23826665348476836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,64,0.00258933338854048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,2560,0.08650044600168864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,5120,0.014200000299347771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,1536,0.005522666705979242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,768,0.006730666591061487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,2048,0.017245334055688646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,7168,0.3417199982537164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,4096,0.012333333492279053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,128,0.02625422179698944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,8192,0.01605688863330417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,5120,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,768,1.2894311481051974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,65536,1.5567946963840062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,64,0.005723555468850666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,768,0.19798133108350965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,16384,0.060416890515221484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,12288,0.03259733319282532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,12288,0.3058648904164632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,6144,1.1368995242648654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,12288,0.12467643949720596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,51200,0.9171262317233615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,4096,0.09908088710572983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,128,0.0064177778032090925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,8192,0.2981742223103841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,8192,0.034834666384591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,6144,0.012433778080675336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,12288,0.932005352444119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,16384,0.0490497781170739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,1536,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,65536,0.3034862147437202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,3584,0.06948444578382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,7168,0.033858666817347206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,3584,0.007677333222495184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,64,0.0069520001610120135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,3072,0.012501333322789935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,1536,0.03886222177081638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,16384,0.1724177731408013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,16384,0.024029332730505202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,51200,0.9605609046088325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,65536,0.3672933313581679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,2048,1.3334524366590712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,1024,0.016689777374267578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,1536,0.15070754951900905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,16384,0.058727112081315785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,5120,0.1827911138534546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,65536,0.8434515529208713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,768,0.008650666309727563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,8192,0.2982444498274061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,64,0.008030222521887885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,128,0.0031866667171319327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,3584,0.00889866633547677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,7168,0.10717333025402492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,12288,0.0699253347184923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,768,0.04690044456058078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,51200,1.2005875905354817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,256,0.0038080000215106537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,768,0.00924533357222875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,3584,0.009703999592198266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,32,0.0074373334646224976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,7168,0.018534221582942538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,128,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,16384,0.20012799898783365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,2560,0.15588088830312094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,32,0.028413332170910303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,7168,0.013373333546850415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,256,0.0037671112351947357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,1024,0.03928177886539035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,1536,0.007961778177155389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,256,0.005559999909665849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,4096,0.007872888611422645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,5120,0.009175111022260454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,2560,0.007901333272457123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,4096,0.13183644082811144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,1536,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,1024,0.003933333274390963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,7168,0.0566506650712755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,12288,2.760792838202583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,2048,0.01754844519827101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,7168,0.015576000014940897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,1536,0.008611555728647444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,512,0.004447111239035924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,5120,0.022125333547592163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,256,0.005972444597217772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,3584,0.009297777381208207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,1536,0.05693688657548693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,2560,1.7005066341824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,512,0.06257066461775038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,256,0.018840889135996502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,6144,0.03010400136311849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,51200,1.1287635167439778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,5120,0.012168889244397482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,16384,3.3520569271511502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,128,0.06565244330300225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,768,0.025610667135980394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,2048,0.07182488838831584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,51200,0.18711466259426543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,7168,0.030534221066368952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,10240,2.1882603963216147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,8192,0.029142220815022785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,3584,0.06460266643100314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,12288,0.6700311236911349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,16384,0.09078400002585517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,1536,0.008245333201355403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,512,0.0038862224254343244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,1536,0.028416888581381902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,3584,0.034712887472576566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,64,0.004469333423508538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,64,0.006743999818960826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,128,0.01974311139848497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,32,0.0032871110985676446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,8192,0.03292799989382426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,32,0.007137777904669444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,4096,0.029772443903817072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,7168,0.012495111260149213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,512,0.03141333328353034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,3584,0.00998044427898195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,6144,0.36173333062065977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,32,0.013032889200581444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,32,0.017010667257838778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,512,0.010006222460005019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,3584,0.026149332523345947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,6144,0.08405866887834336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,32,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,10240,5.227149539523654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,2560,0.012755555411179861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,768,0.0069297779765393995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,51200,0.05602133274078369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,2560,0.015026667051845126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,256,0.005498666730191972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,5120,0.011021333436171213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,5120,0.026709333062171936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,1536,0.012191111014948951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,7168,0.00923644420173433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,1536,0.0046640001237392426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,4096,0.1121751070022583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,1536,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,2048,0.01626044511795044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,256,0.005163555757866966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,1536,0.009467555416954888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,51200,0.07799466451009114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,4096,0.011523555550310345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,6144,0.3981688817342122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,2048,0.005619555711746216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,6144,0.013628444737858243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,512,0.016419554750124615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,768,0.012980444563759698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,5120,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,3072,0.10806310839123195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,256,0.0037928890022966596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,2048,2.9478622012668185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,2560,0.00904444439543618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,12288,0.02029244436158074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,6144,0.22502401140001085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,10240,0.07204711437225342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,2048,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,2048,0.005575110928879843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,2560,0.029623111089070637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,6144,0.020837333467271592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,8192,0.40756267971462673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,6144,0.014096000128322177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,768,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,512,0.05686133437686496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,65536,14.851584540473091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,2048,0.01531644496652815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,512,0.003576888806290097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,12288,0.017460443907313876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,16384,0.010644444160991244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,8192,0.8214453591240777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,128,0.015203555425008139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,5120,0.011668444507651858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,512,0.05391022231843737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,768,0.019127110640207928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,4096,0.008770667016506195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,512,0.03477777706252204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,7168,0.09045688973532783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,256,0.0094355551732911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,51200,0.06311200062433879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,5120,0.006773333168692059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,2048,0.007139555282062954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,5120,0.0122079998254776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,64,0.24738666746351454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,16384,0.47933064566718203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,768,0.011583999627166323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,8192,0.2587413258022732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,1536,0.09201866388320923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,4096,0.013202666408485837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,2048,0.05407199925846524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,16384,0.22911379072401258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,10240,1.6857439676920574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,128,0.016339555382728577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,1024,0.010270222193664974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,2048,0.012330667012267642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,128,0.014139556222491793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,128,0.007354666789372762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,10240,0.028871999846564397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,3072,0.04915822214550442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,1024,0.029727111260096233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,4096,0.011811555259757571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,3072,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,3072,0.030472887886895075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,3584,0.01521511044767168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,3072,5.1252869500054254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,128,0.0035457776652442086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,8192,0.03903822104136149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,12288,0.23703021473354766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,65536,1.0101057688395183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,10240,0.01742755538887448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,3584,0.023031110564867657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,768,0.0035502223504914176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,10240,0.06530044476191203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,51200,0.28049156400892467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,256,0.007106666763623555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,768,0.005772444523043103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,10240,0.028061333629820082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,256,0.011374221907721626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,12288,0.01928711103068458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,7168,0.8704764578077527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,2048,0.016030222177505493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,64,0.006544888847404056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,1024,0.006505777852402792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,1024,0.03230133321550157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,1536,0.0210853334930208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,5120,0.019205333458052743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,6144,0.11126666598849827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,51200,0.06926133235295613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,2560,0.5965893533494737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,8192,0.05313600103060404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,32,0.009039999710188972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,5120,0.01497066683239407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,2560,0.007359111474619971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,12288,0.014654222461912366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,6144,1.5446293089124892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,51200,4.596483442518446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,32,0.006135111053784688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,3584,0.008600888980759515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,51200,0.291723542743259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,1024,0.005061333378156026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,2048,0.00740622232357661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,16384,0.02583733366595374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,8192,0.09563555320103963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,128,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,64,0.003145777723855443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,64,0.0029031110720502008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,128,0.002671111168132888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,7168,0.04519288738568624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,1024,0.8867626720004611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,4096,0.010444444086816577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,6144,0.02739199995994568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,3072,0.48876354429456925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,2048,0.012056888805495368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,5120,0.030288888348473444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,4096,1.2860871420966256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,1536,0.4137590991126166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,6144,0.022138666775491502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,3072,0.054924445019827954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,2048,0.02581600017017788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,8192,0.19836178090837267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,32,0.007246221933099959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,512,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,12288,0.11734755833943684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,4096,0.05093422200944689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,768,0.05359466539488899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,6144,0.1844782167010837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,128,0.004187555362780889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,8192,0.15575910939110651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,2048,0.019277332557572257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,12288,0.033711112207836576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,6144,0.32812534438239205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,1024,0.011704000333944956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,10240,0.06196888950135973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,7168,0.01183377785815133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,4096,0.03814399904674954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,2560,0.009202666580677032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,8192,0.014121777481502958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,768,0.02165422174665663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,512,0.008597333398130205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,3072,0.013956444131003486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,256,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,1024,0.0058133333093590206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,16384,0.8678728739420573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,5120,0.07202666335635714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,512,0.019167110323905945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,65536,0.08354577753278945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,2048,0.006285333385070165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,4096,0.26005956861707896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,768,0.004030222280157937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,65536,0.06022310919231839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,10240,0.06681599881913927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,128,0.0068604445291890045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,256,0.0035546666218174826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,1024,0.00416711096962293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,65536,0.41212799814012313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,4096,0.0243511117166943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,2048,0.01423555612564087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,7168,0.271799988216824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,6144,0.2869173420800103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,2048,0.03461866577466329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,32,0.0041288890772395665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,128,0.19044266806708443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,8192,0.014856888188256158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,5120,0.017521777086787753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,3072,0.010242666635248397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,1536,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,128,0.006701333241330252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,6144,1.222977744208442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,16384,1.0545501708984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,128,0.009768889182143742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,128,0.002700444517864121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,64,0.0634960002369351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,8192,0.12632800473107234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,4096,0.0805386635992262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,2560,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,16384,0.2979857656690809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,1536,0.04836622211668226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,65536,0.06602044237984551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,1536,0.011509333219793109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,512,0.023264888260099623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,6144,0.020302222834693063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,7168,0.017300445172521803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,12288,0.2134782208336724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,2048,0.06325688627031115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,1536,0.014544889330863953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,1024,0.004774222357405556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,12288,0.3563600116305881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,5120,0.017628444565667045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,12288,0.019632889164818656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,512,0.00833777752187517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,2560,0.2640559938218859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,2560,0.009350222018029954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,16384,0.4243884351518419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,3072,0.013256000147925483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,3072,0.011424000064531961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,7168,0.10710844728681777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,7168,0.04743822084532844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,512,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,6144,0.05351110961702135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,4096,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,32,0.002741333304180039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,128,0.007461333440409766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,32,0.004337777694066365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,1536,0.01604711181587643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,51200,0.13307644261254206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,5120,0.010229333407349056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,2048,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,3584,0.8476533359951444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,4096,0.11164532767401801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,128,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,6144,0.017120889491505094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,256,0.0029724445194005966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,51200,1.7350471284654405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,7168,0.2721644507514106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,4096,0.3126800060272217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,51200,0.010218666659461128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,64,0.0029760001020299066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,3584,0.009111999637550777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,1536,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,3072,0.014256000518798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,65536,0.1520062287648519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,5120,0.007146666447321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,16384,0.05948444207509359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,3584,0.014514666464593677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,512,0.07259200016657512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,128,0.005011555635266834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,256,0.004264889078007804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,2048,0.04965955681271023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,1536,0.03943377733230591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,7168,0.03058755397796631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,5120,0.009329777624871995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,12288,0.037779556380377874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,5120,0.16356800662146673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,3584,0.006686222222116258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,2048,0.011167999770906238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,7168,0.008329778081840938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,2048,0.7805937661064996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,768,0.007079111205206976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,12288,0.020303999384244282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,8192,0.21446222729153105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,16384,0.025447110335032146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,512,0.005064888960785336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,7168,0.01364177796575758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,256,0.00815822184085846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,8192,0.024099555280473497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,10240,0.19236089123619926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,16384,0.10462933116488987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,1024,0.0433555543422699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,16384,0.5572969118754069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,256,0.0032462223122517266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,512,0.00638133328821924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,1536,0.015805333852767944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,16384,0.049117333359188504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,3072,0.12547022766537136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,6144,0.010037333601050908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,1536,0.011855111353927188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,6144,0.19376356071896025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,768,0.0058666666348775225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,2048,0.06813600328233507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,3072,0.010580444501505958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,2048,0.01145511120557785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,16384,5.851293351915147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,65536,3.042264090643989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,2560,0.21299289332495797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,16384,1.0250782436794705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,256,0.007915555602974361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,12288,0.03233511249224345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,51200,85.38026936848958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,128,0.00960711141427358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,8192,0.051051557064056396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,2560,0.008539555801285638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,2560,0.009013333254390294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,65536,0.7077680163913302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,768,0.021145777569876775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,4096,0.011028444601429833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,3584,0.006828444285525217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,32,0.003085333233078321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,6144,0.046257777346505054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,16384,0.11329421732160781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,12288,0.020816889074113634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,1536,0.04574844572279188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,1536,0.015999999311235216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,64,0.006728000111050076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,512,0.013062222136391534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,51200,2.055277294582791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,768,0.004811555561092165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,512,0.03791466686460707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,7168,0.03542844454447428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,4096,0.05664622121387058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,2560,0.1732728878657023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,3584,0.025808889004919264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,2048,0.04713244570626152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,3584,0.04461866617202759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,2048,0.04774666825930277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,16384,0.5244613223605686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,1536,0.03577955563863119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,51200,0.1416506634818183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,3584,0.12184533807966445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,16384,0.1950204504860772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,4096,0.06140444676081339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,512,0.03503911031617059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,128,0.0038488888078265717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,256,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,65536,2.351033740573459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,51200,0.1453004413180881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,3072,0.14804711606767443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,3072,0.01607022186120351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,10240,0.016661332713233102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,1536,0.006699555450015598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,16384,0.5127431021796333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,128,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,5120,1.035403569539388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,2048,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,16384,0.011512888802422417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,32,0.0034497777620951333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,16384,0.02277244461907281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,12288,0.006114666660626729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,512,0.04203288753827413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,6144,0.010215111076831818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,65536,0.24569866392347547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,12288,0.04521422253714668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,7168,0.021168889270888433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,64,0.003212444484233856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,1536,0.03205422229237027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,3584,0.11954132715861003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,64,0.0033164444482988785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,16384,0.8460488849216037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,2560,0.007185777856243982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,256,0.0036115555299652945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,1536,0.006485333459244833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,5120,2.4134754604763455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,128,0.02427022159099579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,65536,0.09309511052237616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,3072,0.8080106841193305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,65536,0.2538391219245063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,8192,0.012162666353914471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,7168,0.13290666209326849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,2048,0.04928000105751885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,64,0.013895110951529609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,1536,0.012696888711717395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,16384,0.024884444144037034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,8192,0.040349334478378296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,12288,0.04250577754444546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,3072,0.020115555988417733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,3072,0.03892711136076186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,65536,0.10110221968756782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,6144,0.04768622252676222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,3584,0.6652853224012587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,128,0.005997333261701796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,4096,0.06899822420544095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,256,0.06688088840908475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,512,0.00406577769252989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,16384,0.03500533435079787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,1536,0.010543111297819348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,16384,0.520311090681288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,16384,0.10082577996783787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,51200,9.772833930121527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,4096,0.014523555835088095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,5120,0.039149334033330284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,3072,0.02609511050913069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,2048,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,6144,0.030008001459969416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,10240,0.09933955801857842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,6144,0.010553778045707278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,1024,0.008727110922336578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,2048,0.006667555620272954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,4096,0.05568089087804159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,128,0.0037288889288902283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,12288,0.3004488945007324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,12288,0.035637332333458796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,1536,0.0911288857460022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,6144,0.25373954243130153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,6144,0.08979199992285834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,2048,0.011167999770906238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,10240,0.510108470916748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,51200,1.0175715552435982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,1536,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,32,0.00815555536084705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,64,0.03820266657405429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,128,0.031876444816589355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,512,0.010058666268984476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,12288,3.920653449164497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,7168,0.06090044313006931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,1024,0.009127111070685917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,6144,0.030815998713175457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,4096,0.03185155656602647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,51200,0.05521600113974678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,1024,0.01017599966790941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,128,0.026660444007979497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,1024,0.006797333144479328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,1024,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,128,0.0034764446318149567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,2560,0.008286221987671321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,16384,0.8149440023634168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,2560,0.29730576939053005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,10240,0.08171111345291138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,6144,0.010955555571450127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,8192,0.01664888858795166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,3584,0.010306666294733683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,10240,0.18267199728224012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,3584,0.02513866623242696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,1536,0.005673777725961473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,3584,0.010233778092596266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,65536,0.15662754906548396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,16384,0.026872888207435608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,12288,0.04840622345606486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,51200,0.04761244522200691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,256,0.14495910538567439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,32,0.22374844551086426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,10240,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,128,0.0031519999934567344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,256,0.0052426668504873914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,256,0.010959111154079437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,4096,0.007881777981917063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,7168,1.0952000088161893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,1536,0.012730666332774691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,2048,0.016210666961140104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,12288,0.020761777957280476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,5120,0.02001688877741496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,10240,0.020872889293564692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,4096,0.010983111129866706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,32,0.006769777586062749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,6144,0.011039111349317761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,4096,0.008262222011884054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,12288,0.1829448938369751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,768,0.008559111091825698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,16384,0.27043554517957896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,128,0.044960889551374644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,256,0.00889511075284746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,1024,0.01685333251953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,1024,0.009112888740168678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,2560,0.08953333563274807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,7168,0.04340089029735989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,1024,0.005227555417352253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,4096,0.022114665971861944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,4096,0.012160888976520963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,65536,1.2431821823120117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,768,0.005984888722499211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,512,0.005943999936183293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,3072,0.14275999863942465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,12288,0.12416533629099528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,1024,0.018808888064490426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,1024,0.027296889159414504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,512,0.005976888868543837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,7168,0.020752888586786058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,768,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,4096,0.11586666107177734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,5120,0.21195199754503039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,128,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,512,0.005187555733654234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,4096,0.20764622423383924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,7168,0.057378669579823814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,2048,0.09884355465571086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,3584,0.06714933448367648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,51200,0.05973155630959404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,768,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,10240,0.07745777898364596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,2560,0.1167715522978041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,10240,0.048935112026002675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,51200,0.06683466831843059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,7168,0.012423111332787408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,8192,0.026867555247412786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,128,0.009933333430025313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,1536,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,2048,0.0074106670088238185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,12288,0.926886240641276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,3584,0.01761777698993683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,3072,0.04103022151523166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,768,0.005264889034960005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,32,0.002565333412753211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,4096,0.009296000003814697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,51200,0.643846246931288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,16384,0.053256889184316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,8192,0.01640533407529195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,12288,0.13202577167087132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,3072,0.011695110963450538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,8192,4.789526197645399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,8192,0.1447733375761244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,3584,0.17476888497670492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,8192,0.016042666302786935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,16384,0.06885600090026855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,512,0.005224888937340842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,10240,1.6904462178548176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,4096,1.0153919855753581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,2560,0.008223999705579545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,32,0.004292444636424382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,51200,0.045181334018707275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,5120,0.009901333186361525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,8192,0.01607733302646213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,2560,0.01162577751610014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,1024,0.07591733005311754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,512,0.06285244226455688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,768,0.003998222036494149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,64,0.003175111073586676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,65536,3.5554612477620444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,12288,0.10478222370147705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,16384,0.04248799880345663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,2560,0.03424800104565091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,32,0.005778666585683823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,4096,0.010819555156760745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,3584,0.01825777855184343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,512,0.03201244605912103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,512,0.005564444594913059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,65536,0.19181511137220594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,4096,0.027679110566775005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,3072,0.0075866662793689305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,4096,0.16481510798136392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,5120,0.023942222197850544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,512,0.021324444148275588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,3072,0.06902577479680379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,128,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,128,0.005889777921968036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,256,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,512,0.008330666356616551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,65536,12.297995673285591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,7168,0.034903112385008074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,32,0.0026373333401150177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,12288,0.7920079761081271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,2560,0.04865511258443197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,3584,0.3584551016489665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,65536,0.8100453482733833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,64,0.0029208889851967492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,768,0.024359999431504145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,2560,0.008546666966544257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,3584,0.010520888699425591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,12288,0.31579287846883136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,8192,0.009334222310119206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,1536,0.011797332929240333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,32,0.0027662221756246355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,1536,0.011357333097192975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,5120,0.008718222379684448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,10240,0.05356088942951626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,256,0.004558222161398994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,10240,0.016386666231685214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,3072,0.006775110960006714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,2048,0.04440977838304308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,256,0.007486222518814935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,128,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,51200,2.2384605407714844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,512,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,1024,0.0052897776994440295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,16384,0.02645155621899499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,4096,0.10370577706231011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,768,0.010874666273593903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,512,0.008076444268226624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,7168,0.20645422405666777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,5120,0.1538177728652954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,1024,0.006058666855096817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,7168,0.13467378086513943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,3072,0.030840890275107488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,64,0.005340444544951121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,7168,0.006832888970772426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,7168,0.008107555409272512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,512,0.5474328994750977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,768,0.05836711327234904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,3072,0.42423023117913145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,4096,0.25683644082811147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,128,0.04496177699830797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,7168,0.10083378023571438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,64,0.002936888900068071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,12288,0.007744000189834171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,1024,0.0058240000572469496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,768,0.0277804434299469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,32,0.0027928888383838865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,12288,0.018920888503392536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,64,0.005513777749405966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,16384,4.876984066433377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,256,0.023762666516833838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,51200,0.3076293203565809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,6144,0.00674222202764617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,3072,0.009256000320116678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,3072,0.018333334061834548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,5120,0.01570311188697815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,128,0.014707555373509726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,51200,0.07493422428766887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,8192,0.010332444475756751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,16384,0.022123555342356365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,1024,0.021103999680942957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,12288,0.018599111172888014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,10240,0.19677066802978516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,2048,0.05432888865470886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,1536,0.004659555438492033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,16384,0.13680888546837702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,12288,0.05377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,51200,0.009566222627957663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,2560,0.019127999742825825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,768,0.04346222347683377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,64,0.002801777794957161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,32,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,64,0.002622222320901023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,7168,0.009365333451165093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,6144,0.10254044665230645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,2560,0.05136000116666158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,32,0.005163555757866966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,3072,0.0074755557709270054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,5120,0.008647999829716152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,256,0.003109333415826162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,64,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,4096,0.020908445119857788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,51200,2.162439982096354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,512,0.030163556337356567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,1536,0.01480977733929952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,3584,0.17442844973670113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,768,0.00981155534585317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,4096,0.16121955712636313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,512,0.0248355550898446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,1024,0.005790222022268508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,16384,0.030983110268910725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,1536,0.008771555291281806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,3072,0.024235554867320593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,51200,0.4420319928063287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,3584,0.00998222248421775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,3072,0.08632089032067193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,7168,0.08019200298521253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,2560,0.008037333687146505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,256,0.012542222109105853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,64,0.004074666649103165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,512,0.009111999637550777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,256,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,10240,0.017439999514155917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,5120,0.03104089034928216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,7168,0.05340622199906243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,768,0.00360000009338061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,512,0.004777777940034866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,10240,0.013405332962671915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,2560,0.009088888764381409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,1536,0.0325893329249488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,6144,0.009568000005351173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,1024,0.004536889079544279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,512,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,128,0.009390222529570261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,51200,4.44499037000868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,64,0.011957333319716983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,12288,0.4719129138522678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,3584,0.012917333179050021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,32,0.004767111192146937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,3072,0.012558221817016602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,3584,0.10398399829864502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,512,0.00527555536892679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,8192,0.07085244523154365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,1024,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,6144,0.07016977998945448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,16384,4.237439897325304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,768,0.026680888401137456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,256,0.020575111111005146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,2560,0.01257955531279246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,1024,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,51200,1.5929092831081813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,16384,2.022768020629883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,128,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,256,0.0035217776894569397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,16384,0.04256711072391934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,128,0.05240799983342489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,1024,0.015159999330838522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,1024,0.00903555585278405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,3072,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,128,0.011884444289737277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,1536,0.01832355558872223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,4096,0.06637777884801228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,51200,1.7345369127061632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,2048,0.06025866667429606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,768,0.006065777606434292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,16384,0.06627021895514594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,6144,0.12038311693403457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,2560,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,16384,0.039493332306543984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,64,0.011362666885058085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,4096,0.013800889253616333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,12288,3.4972337086995444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,1536,0.013409777647919126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,5120,0.02051111062367757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,5120,0.29282932811313206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,16384,0.24766132566663954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,10240,0.0244195560614268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,2048,0.008614222208658854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,512,0.10444888803693984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,6144,0.044807109567854136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,512,0.01613333324591319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,6144,0.0148062225845125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,7168,0.014263111684057446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,3072,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,2048,0.7231600019666883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,2048,0.0064737776087390045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,5120,0.051071110698911876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,7168,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,7168,0.012873777912722694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,768,0.037864890363481306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,128,0.0052266667286554975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,32,0.002578666640652551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,51200,0.36993779076470273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,1536,0.02604266670015123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,128,0.002696000039577484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,12288,0.03256622287962172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,2048,0.014815110299322339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,8192,0.05344088872273763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,10240,0.013012444807423485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,16384,0.02481422159406874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,32,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,8192,0.01181422256761127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,128,0.2191173368030124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,2048,0.2654942141638862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,256,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,256,0.005666666560702854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,1536,0.01349244432316886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,6144,0.015241778559154935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,51200,0.09886311160193549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,3072,0.05335466729270088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,256,0.007048000064161088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,1024,0.008742222355471717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,128,0.02216088937388526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,3584,0.010031110710567897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,1024,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,3584,0.03817688756518894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,2560,0.010306666294733683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,16384,0.05366666780577766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,65536,0.21002311176723906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,128,0.005917333480384614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,2560,0.011327111058764987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,7168,0.09899022181828816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,4096,0.009220444493823582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,3584,0.019448889626397025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,8192,0.029306666718588933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,7168,0.01738666660255856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,6144,0.03446933295991685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,1536,0.006488889041874144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,16384,0.016647110382715862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,64,0.025664000047577754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,3072,0.07551022370656331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,768,0.5760062005784776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,3072,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,6144,0.022456000248591106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,512,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,4096,0.0735786689652337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,5120,0.012328888807031842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,3072,0.21112799644470215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,64,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,51200,0.3728613323635525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,16384,0.046810666720072426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,768,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,3584,0.05478044350941976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,6144,0.04447466797298855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,65536,7.473264058430989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,6144,0.014300444059901767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,8192,0.33702399995591903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,1024,0.1149946716096666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,7168,0.8389911121792264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,64,0.0038213332494099936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,256,0.01391377713945177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,3584,0.021271111236678228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,4096,0.14311556021372476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,1536,0.0047315553658538395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,4096,0.030702223380406696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,768,0.04001244571473863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,65536,0.032263110081354775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,512,0.008736888567606607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,3072,0.0069653333889113525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,128,0.02052177819940779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,1024,0.007000000112586551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,6144,1.2943599489000108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,65536,2.030544916788737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,65536,0.10102133618460761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,128,0.0042151109211974675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,256,0.07493422428766887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,16384,0.05188444587919447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,1536,0.0046728890803125174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,1024,0.006090666684839461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,7168,0.0990755558013916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,51200,0.4954569074842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,10240,0.01808888879087236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,5120,0.0931475559870402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,7168,0.1325297753016154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,10240,0.018384888768196106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,3072,10.940632290310331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,1024,0.15091644393073186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,4096,0.012900444368521372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,51200,0.9335556030273438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,16384,0.48988087972005206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,6144,0.29583999845716685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,4096,0.007054222126801808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,1536,0.008812444077597724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,12288,0.03568177753024631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,12288,0.39302221934000653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,12288,0.19965867201487222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,64,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,7168,0.007055111229419708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,1536,0.17751733462015787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,512,0.011824000212881299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,16384,0.6335937711927626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,128,0.009005332986513773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,8192,0.03326133224699233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,16384,0.025088000628683303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,10240,0.09739466508229573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,512,0.021672000487645466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,256,0.2082986699210273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,32,0.11272177431318496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,3072,0.07100533114539252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,6144,0.21138666735755074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,3584,0.013988445202509562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,12288,0.01566488875283135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,16384,0.028839111328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,65536,0.08212800158394708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,768,0.03497599893146091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,51200,0.7877324422200521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,3072,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,51200,0.22370844417148164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,12288,0.6466817855834961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,64,0.010314666562610203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,51200,9.87542470296224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,512,0.025759999950726826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,512,0.04105777872933282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,256,0.007736889024575551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,3072,0.013649777405791812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,64,0.002887111157178879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,10240,1.7368141810099285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,1024,0.9796489079793295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,8192,0.09087910917070176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,6144,0.01850133306450314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,65536,0.2883315616183811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,10240,2.5785039265950522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,2048,0.09228533506393433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,3072,0.02069155540731218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,7168,0.012960889273219638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,65536,3.7701636420355906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,2048,0.047447111871507436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,1536,0.009071110851234859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,1024,0.010094222095277574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,32,0.0037875556283526947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,6144,0.017624000708262127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,16384,0.0748071140713162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,51200,0.06182577874925402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,10240,0.05299377772543165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,65536,0.07434133026334974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,5120,0.020845333735148113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,2048,0.005366222312053044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,51200,0.06234311395221286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,32,0.0052266667286554975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,3072,0.09724799791971843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,3584,0.019184889064894784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,768,0.006301333506902059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,10240,0.038535998927222356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,768,0.0289066665702396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,7168,0.25363998942905003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,8192,0.07171110974417792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,512,0.011552888486120435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,64,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,12288,0.12337956163618301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,4096,2.0700115627712674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,16384,57.7145250108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,16384,0.02167555524243249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,12288,0.029284444120195176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,2048,0.01738222274515364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,256,0.006291555447710886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,2560,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,32,0.005583111196756363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,6144,0.1580959955851237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,128,0.04702577657169766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,3584,0.007483555210961237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,6144,0.017491555876202054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,4096,0.009966221948464712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,128,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,3584,0.05524355504247877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,5120,0.2910382217831082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,6144,0.01311822235584259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,64,0.005009777843952179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,2048,0.004996444616052839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,10240,0.016217778126398723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,65536,0.1560702191458808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,3584,0.012098666694429187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,4096,0.010441777606805166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,65536,2.7664754655626087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,1024,0.007059555914666917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,2560,0.2778080039554172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,768,0.003991111285156674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,3584,0.01236177815331353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,12288,0.07961333460277982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,8192,0.01018044435315662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,5120,0.011226666470368704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,256,0.05645955271191067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,51200,0.3145048883226183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,64,0.007619555625650618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,64,0.027912888262006972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,2048,1.0672231250339084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,3584,0.00941422250535753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,12288,0.32833067576090497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,64,0.006877333339717653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,65536,0.3315822283426921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,6144,0.011042666931947073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,2560,0.02647022240691715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,1024,0.036990222003724836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,1536,0.010631110933091907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,51200,0.3735822306738959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,51200,50.89281548394097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,64,0.0081404447555542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,4096,0.010182222558392419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,65536,32.064551459418404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,2048,0.018631999691327412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,3584,0.012019555601808759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,64,0.005535999933878581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,8192,0.013573333621025085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,65536,0.292592896355523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,65536,0.47468890084160703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,3072,0.01572533282968733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,768,0.01057866629627016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,128,0.004867555366622077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,12288,0.016381333271662395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,3584,0.06682044267654419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,51200,1.4489190843370225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,10240,0.11527289284600152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,2560,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,16384,0.04331199990378486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,16384,0.2495182090335422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,8192,0.21754577424791124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,1536,0.006443555570311016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,6144,0.007844444778230455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,4096,0.011020444333553314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,6144,0.04099911120202806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,32,0.0026737778551048706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,64,0.07608622312545776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,6144,0.015224888920783997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,5120,0.014255111416180929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,64,0.005054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,12288,0.00997511131895913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,64,0.0026444445053736367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,10240,0.2645137839847141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,5120,0.013088000317414602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,65536,2.1265830993652344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,10240,0.027642667293548584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,51200,2.582711113823785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,4096,1.06513426038954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,65536,0.016569778323173523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,5120,0.37925065888298887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,2048,0.007500444849332173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,32,0.006623111251327727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,2560,0.6047360102335612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,128,0.016297777493794758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,1536,0.023025777604844835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,64,0.006595555692911148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,2048,0.08472800254821777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,32,0.002970666728085942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,4096,0.05855466922124227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,2560,0.05260622170236376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,64,0.0038844446341196695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,128,0.0033599999215867785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,12288,0.018274666534529794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,10240,0.0327413347032335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,1536,0.008627555436558193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,128,0.008560889297061497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,1536,0.004661333229806689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,51200,0.19425866338941786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,5120,0.07767999834484525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,10240,0.019985778464211356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,768,0.003961777935425441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,3584,0.055192000336117215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,2560,0.018106665876176622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,1024,0.006931555353932911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,5120,0.16655377546946207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,6144,0.2281200091044108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,1024,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,128,0.0029893333299292457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,51200,0.7030231157938639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,128,0.0030577776746617425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,51200,0.7696702215406629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,16384,0.008808000220192803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,5120,0.13569333818223742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,8192,3.826202816433377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,6144,0.013849777479966482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,3584,0.31479912334018284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,5120,0.10320978032218085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,12288,0.23916800816853842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,2560,0.010465777582592435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,65536,1.791428460015191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,8192,0.02040177749262916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,5120,0.021659556362364028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,64,0.002926222152180142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,8192,0.06760533650716145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,4096,0.012970666918489667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,4096,0.011746666497654386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,10240,0.021690666675567627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,32,0.0034577778230110803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,2560,0.00996800015370051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,8192,0.027839110957251653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,51200,0.19469955232408312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,6144,0.06337510877185397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,8192,3.1176711188422312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,1024,0.26939466264512807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,10240,0.040778666734695435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,3584,0.012826667063766055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,512,0.006696888970004187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,1536,0.060138669278886586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,8192,0.06528266933229235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,2560,0.0282008879714542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,256,0.005803555664088991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,64,0.005390222287840313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,7168,0.1816471152835422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,3584,0.02996088729964362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,16384,0.021617778473430212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,8192,1.530069351196289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,10240,0.6847084363301595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,3072,0.023510222633679707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,4096,0.5376737912495931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,768,0.0046942221621672315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,512,0.05096800128618876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,128,0.00508088908261723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,512,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,64,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,128,0.004839999808205498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,512,0.01715288890732659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,32,0.007345777418878343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,7168,0.023317333724763658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,3072,0.09383289019266765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,2560,0.011232888533009423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,4096,0.006623111251327727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,64,0.012705778082211813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,16384,1.3288888931274414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,512,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,6144,0.020399111840460036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,1536,0.08593866560194229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,5120,0.17527733908759224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,1536,0.019863110449579027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,16384,0.4190017647213406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,2560,0.0199306673473782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,65536,0.07508889171812269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,6144,0.013305777476893531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,32,0.0030213333666324615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,6144,0.020776000287797716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,3584,0.020346666375796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,51200,0.08660799927181667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,3072,0.013853333062595792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,51200,0.193704883257548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,1024,0.05312088794178433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,51200,1.8747920989990234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,512,0.011619555453459421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,5120,0.21961332692040336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,65536,0.2126577827665541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,768,0.013004444539546967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,51200,0.018594667315483093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,32,0.006686222222116258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,6144,0.04383644461631775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,51200,0.3275715510050456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,768,0.015138667490747241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,12288,0.019307555423842538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,6144,0.03085600005255805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,6144,0.026152888933817547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,12288,0.05307022068235609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,64,0.01074399964676963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,65536,0.3467093308766683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,2560,0.08433510859807332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,2048,0.007176888485749562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,3072,0.00904177791542477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,5120,0.023041778140597876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,10240,0.24421244197421602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,3584,0.011623111036088733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,3072,0.01332266628742218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,1024,0.008739555875460306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,2048,0.011746666497654386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,7168,0.009132444030708736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,5120,0.009010666774378883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,65536,1.0917492972479927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,2560,0.011046222514576383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,8192,0.016361777981122334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,2048,0.013714666995737286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,2048,0.04055555661519369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,2560,1.0258329179551866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,65536,0.14442400137583414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,2560,0.007663110891977946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,64,0.119851549466451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,8192,0.12933599948883057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,512,0.005471999860472149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,128,0.043185777134365506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,16384,0.010400888820489248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,8192,0.13989244567023382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,16384,0.08427466948827107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,512,0.0064399999876817065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,768,0.04256622327698601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,65536,0.1629128853480021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,3584,0.008691555923885768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,16384,0.09349066681332058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,6144,0.02387822171052297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,10240,0.0315048893292745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,5120,0.11070133580101861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,256,0.019754666421148512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,2048,0.015949333707491558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,65536,0.1416248877843221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,5120,0.014241778188281588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,768,0.03663999835650126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,51200,0.04538133409288195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,1536,0.009520000053776635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,64,0.0028719999310043124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,5120,0.3075057665506999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,5120,0.053207112683190234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,16384,0.9350631501939562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,51200,1.567096922132704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,1536,0.10205422507392035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,1024,0.008398222426573435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,7168,0.0074782222509384155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,512,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,3584,0.00962488849957784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,512,0.003955555458863576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,16384,0.05862310859892103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,16384,0.5320977634853786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,3072,0.014908444550302295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,2560,0.035975111855400935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,10240,0.01903555625014835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,3072,0.05192800031767952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,51200,0.7819813092549642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,8192,0.02203911046187083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,1536,0.03306311037805345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,8192,0.2930826610989041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,51200,0.07549066676033868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,2560,0.04820266697141859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,3584,0.34316444396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,512,0.004948444250557157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,16384,0.05326399869389004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,768,0.008958222137557136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,256,0.05431288811895582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,7168,0.045336001449161105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,128,0.0028648889727062653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,6144,0.1143066618177626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,768,0.007081777685218387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,32,0.004951999833186467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,3072,0.019966221517986722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,3072,0.013817778064144982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,256,0.0031893334041039148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,16384,0.051948444710837476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,256,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,3072,0.2754266791873508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,12288,0.22175021966298422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,4096,0.016903999778959487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,3584,0.05740888913472494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,51200,18.910898844401043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,12288,0.04167644513977898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,6144,0.21116888523101807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,12288,0.1859431134329902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,768,0.003726222034957674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,2048,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,6144,0.037639998727374606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,3584,0.06875466638141207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,2048,0.01402844488620758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,51200,0.591203530629476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,3072,0.012448000411192576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,16384,0.041512889994515315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,64,0.00398577791121271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,7168,0.025813332862324182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,64,0.03501155641343858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,2560,0.00889244427283605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,8192,0.26131288210550946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,16384,0.1379457712173462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,2048,0.061216890811920166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,8192,0.07054489188724093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,16384,1.7340497970581055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,3072,0.015951999359660678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,3072,0.41321778297424316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,512,0.004498666773239772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,16384,0.3456248972151015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,7168,0.14650932947794595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,768,0.004569777597983678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,16384,0.0736044446627299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,1536,0.006958222223652734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,768,0.005599110904667113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,256,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,10240,0.03851200143496195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,64,0.005103999955786599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,128,0.0027146666414207886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,768,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,1536,0.004691555682155821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,6144,0.054916444751951426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,1536,0.015855999456511605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,5120,0.009170666337013245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,768,0.014228444960382251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,512,0.007903999752468532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,3072,0.008160889148712158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,256,0.004068444586462445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,768,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,768,0.5132444699605306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,2048,0.00776800016562144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,2560,0.02496444516711765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,1536,0.04706933432155185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,16384,0.0222746663623386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,128,0.008347555167145198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,7168,0.44908444086710614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,1024,0.20061422718895805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,512,0.005158222383923001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,768,0.0038888889054457345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,256,0.05561333232455783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,1024,0.01962222158908844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,16384,0.6150222354465061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,51200,0.07495822509129842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,65536,0.1421546671125624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,128,0.009165333376990424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,3072,0.013140444126394061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,10240,0.023668444818920557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,5120,0.23499377568562826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,16384,0.08031022548675537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,256,0.0033413333197434745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,6144,0.013176889055305056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,512,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,32,0.0673324465751648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,256,0.003170666595300039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,4096,0.008303999900817871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,16384,0.08759022421307033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,256,1.6705395380655925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,3584,0.021708443760871887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,3584,0.013246222502655454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,5120,0.20225955380333793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,32,0.0029768887907266617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,32,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,16384,0.007732444339328342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,65536,0.14752711190117732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,2048,0.007065777977307637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,7168,0.1750097804599338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,51200,0.07676888836754693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,16384,0.0404426654179891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,10240,0.08127822478612264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,3584,0.012368888490729861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,256,0.006615110983451207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,3584,0.013632888595263163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,1024,0.030576000610987347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,51200,0.19552800390455458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,64,0.009263111485375298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,6144,0.7775920232137045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,2560,0.009159111314349705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,16384,0.025966222087542217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,1024,0.004884444591071871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,5120,0.01639111174477471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,2560,0.17502489354875353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,512,0.004250666747490565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,7168,0.01182755579551061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,32,0.005469333380460739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,2560,0.010203555226325989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,1024,0.018228444788191054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,128,0.019828443725903828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,128,0.0038480001191298165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,2048,0.12403733200497097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,3072,0.014551111393504672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,7168,0.1627146667904324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,16384,0.1100542214181688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,12288,1.0146257612440321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,1024,0.05507022142410278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,512,0.007156444092591603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,2048,0.01128888875246048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,6144,0.013053333593739403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,8192,0.10554399755265977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,768,0.006161777923504512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,3584,0.011130666567219628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,10240,0.15687199433644614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,10240,0.018067555295096505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,768,0.004944889081848992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,3584,0.02846755584081014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,7168,0.44683202107747394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,6144,0.01620533400111728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,65536,0.08479555447896321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,51200,1.1519591013590496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,512,0.007620444728268518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,8192,0.09068889088100857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,1536,0.012615111139085559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,4096,0.1466159953011407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,51200,0.10043822394476996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,1024,0.00638133328821924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,1536,0.016025778320100572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,128,0.0041786668201287585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,10240,0.027485332555241052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,6144,0.3189217779371473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,1024,0.4582542313469781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,5120,0.2330231136745877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,32,0.014281777871979607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,64,0.015980444020695157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,5120,1.1835963990953233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,1536,0.06863733132680257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,3584,0.06965421968036227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,51200,0.6848693423800998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,4096,0.01565511193540361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,1024,0.020752888586786058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,16384,0.021002666817771062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,65536,0.6304968727959527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,6144,0.30443021986219615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,2560,0.00777599960565567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,12288,0.31867021984524196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,1024,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,2560,0.010981332924630908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,128,0.003924444317817688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,512,0.005780444376998478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,7168,0.09122577640745376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,256,0.009381333159075843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,12288,1.6451706356472438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,12288,0.014203555054134794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,1024,0.012327111429638334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,128,0.011492444409264458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,1024,0.054674665133158364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,10240,0.022314666046036616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,65536,0.1914737754397922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,65536,4.1904958089192705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,2048,0.01296266665061315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,64,0.003531555748648114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,6144,0.029338667790095013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,128,0.00305244450767835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,256,0.023175110419591267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,1024,0.014878221684032016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,5120,0.007496889266702864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,65536,0.10603733195198907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,32,0.01977688901954227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,51200,1.4353200064765081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,2048,0.04007733199331496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,3584,0.012647999657524956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,2048,0.015200000670221118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,1536,0.013935110635227628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,512,0.06469777557584974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,2048,0.0090515555606948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,4096,0.029570665624406602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,65536,5.343326992458767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,2048,0.03267822331852383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,4096,0.010679110884666443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,256,0.019747555255889893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,2560,0.023391111029518977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,3072,5.441416846381293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,12288,0.021021333005693223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,5120,0.013061333033773633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,16384,0.07715733183754815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,7168,0.04773333337571886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,10240,0.018352000249756705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,1024,0.00907555553648207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,64,0.012459555433856117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,7168,0.013205333716339536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,10240,0.06535377767350939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,65536,0.24974934260050455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,2560,0.01552977826860216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,3072,0.011879111329714457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,1024,0.005483555710977978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,512,0.0037546666959921518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,32,0.005097777893145879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,768,0.004855999930037392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,65536,0.21323733859592012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,1536,0.007227555745177799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,51200,0.18175644344753691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,768,0.005290666802061929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,10240,0.013732444081041547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,128,0.004383999854326248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,3072,0.036907556984159685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,3584,0.008424888882372115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,256,0.00425155543618732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,10240,0.019606222709019978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,6144,0.01795288920402527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,512,0.004187555362780889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,2048,0.009679999616410997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,128,0.006800000038411882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,512,0.0070186663005087115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,7168,0.01445777714252472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,3072,0.017839110559887357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,51200,1.3530444039238825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,64,0.006510222123728857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,7168,0.02892177634769016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,6144,0.01588711142539978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,6144,0.027170666389995154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,7168,0.1468506654103597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,3072,0.017812444104088675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,256,0.007848000360859765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,8192,0.04052977760632833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,8192,0.037631998459498085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,2560,0.013625777430004544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,32,0.005037333402368757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,768,0.0921964446703593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,4096,0.011691555380821228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,256,0.018223111828168232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,5120,0.02386666668785943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,7168,0.1822213331858317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,7168,0.20851288901435006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,3072,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,128,0.028063111835055884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,7168,0.5993208885192871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,10240,0.064337776766883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,1024,0.017719111508793302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,768,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,6144,0.010864888628323873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,1536,0.008390222158696916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,3584,0.00979733301533593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,1536,0.2649368974897596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,512,0.006062222023804982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,10240,0.4315484364827474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,32,0.05312888820966085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,3072,0.04563822348912557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,768,0.004090666770935059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,2048,0.2019431061214871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,65536,0.24109689394632974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,12288,0.4105217721727159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,3584,1.1492044660780165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,768,0.0637048880259196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,1536,0.006223111102978389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,2560,0.02700088918209076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,768,0.005168888717889786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,2048,0.09637333287133111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,10240,0.04677599999639723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,4096,0.009537777966923183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,2560,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,512,0.0053280000057485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,16384,0.027299554811583623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,5120,0.01837688850031959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,512,0.0090844440791342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,32,0.008917333351241218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,65536,0.23318134413825142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,12288,0.029198222690158423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,7168,0.01594044433699714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,65536,0.05017422305213081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,128,0.011772444678677453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,65536,0.5404959784613715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,768,0.007833778030342525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,51200,0.06506844361623128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,10240,0.009474666582213508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,6144,0.016556445095274184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,8192,0.29206400447421604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,2560,0.025684444440735713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,64,0.006893333461549547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,12288,0.020748444729381137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,1024,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,3584,0.015970667203267414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,4096,0.12368444601694743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,1536,0.02569777766863505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,512,0.007115555306275685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,8192,4.050859663221571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,3584,0.012803555362754397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,8192,7.5542712741427955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,128,0.07082933187484741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,51200,0.5989039738972982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,2048,0.012423111332787408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,256,0.03363644414477878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,3072,0.11566578017340766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,512,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,3072,0.016968000266287062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,512,0.004590222405062782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,4096,0.017478222648302715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,128,0.00553688903649648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,256,0.043192889955308705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,3072,0.05039644572469923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,10240,0.016656888855828177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,2560,0.019519999623298645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,5120,0.011764444410800934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,5120,0.10914755529827541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,2560,0.05394400159517924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,3072,0.06872800323698255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,3072,0.05517333414819506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,64,0.0029253332565228143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,128,0.09020800060696071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,2048,0.07769066757626004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,6144,0.008246222303973304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,16384,0.037618666887283325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,10240,0.33272356457180446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,32,0.006544888847404056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,12288,0.008199999729792276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,8192,0.10498844252692328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,256,0.002995555599530538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,51200,0.9754169252183703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,65536,0.14841066466437444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,3584,0.009330666727489894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,2048,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,1024,0.006095110956165526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,512,0.005508444375462002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,64,0.007822222179836696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,256,0.004311999926964442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,512,0.01854044364558326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,2048,0.15913155343797472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,768,0.03453777896033393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,32,0.00961866643693712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,6144,0.4819013277689616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,2560,0.013077333569526672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,16384,0.23114932907952201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,1024,0.021288888321982488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,12288,0.22402665350172254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,1024,0.004072000169091755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,5120,0.07406044668621488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,8192,0.015595555305480957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,2560,0.007111999723646376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,3584,0.04134488768047757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,10240,0.02028622229894002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,64,0.0027111110587914786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,3072,1.6084595786200628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,8192,0.0184879998366038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,12288,0.029750221305423315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,4096,0.016688000824716356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,65536,0.151055998272366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,2048,0.06778933604558308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,6144,0.2018044392267863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,3584,0.3709626727634006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,10240,0.06399289104673597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,65536,1.5030480490790472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,3584,0.018106665876176622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,1024,0.006928000185224745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,768,0.03541600041919284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,8192,0.7530337969462076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,5120,0.2168355517917209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,1024,0.007072889142566257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,51200,0.06189155578613281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,51200,0.09688978062735663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,16384,1.2343449062771268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,128,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,3584,0.009500444763236577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,2560,0.0075395554304122925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,16384,0.6517564455668131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,1024,0.08614933490753174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,128,0.01696266730626424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,65536,0.0503039989206526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,16384,2.5405794779459634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,64,0.00960177762640847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,8192,0.02288888891537984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,51200,0.1088835530810886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,4096,0.020755555894639757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,51200,0.10161422358618842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,512,0.01706755492422316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,3584,0.019871999820073444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,3072,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,2048,0.33365334404839414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,8192,0.02035999960369534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,12288,0.4134675661722819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,10240,0.01622488929165734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,128,0.0029297777348094513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,768,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,65536,0.11281600263383652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,6144,0.020452444752057392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,3584,0.03772977656788296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,1024,0.04184266593721178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,2048,0.005306666509972678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,6144,0.05604088968700833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,256,0.005346666607591841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,51200,0.18371110492282441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,51200,1.3823955323961046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,5120,0.02881511052449544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,256,0.005222222043408289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,3584,0.11307556099361843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,8192,0.2637653350830078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,7168,0.010890666809346942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,4096,0.011774222056070963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,768,0.018020444446139865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,2048,0.010863111250930362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,4096,0.1098853349685669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,3072,0.07460711399714152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,2560,0.00999111102686988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,4096,0.5944657855563694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,4096,0.0872560011015998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,51200,4.047440846761067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,6144,0.26673245429992676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,2048,0.09413244326909383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,3072,0.27835022078620064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,51200,2.5946390363905167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,65536,18.593212551540798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,65536,0.504503091176351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,128,0.004007110993067424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,65536,0.2053635517756144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,2560,0.07192889187071058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,768,0.035876442988713585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,16384,0.10049155685636733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,512,0.019648889700571697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,256,0.02422133253680335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,2048,0.01351288871632682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,3072,0.012947555217478009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,5120,0.026480888326962788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,128,0.0040960001448790235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,51200,0.4877155621846517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,768,0.006956444846259222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,1536,0.014097778333557976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,8192,0.10120533572302924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,1536,0.004912000149488449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,256,0.003456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,5120,0.03911022345225016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,3584,0.007326222128338284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,64,0.002499555548032125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,32,0.021215111017227173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,64,0.0026488889836602737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,1024,0.04412000046836006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,128,0.01348355578051673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,1536,0.007370666497283512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,8192,0.03801511062516107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,12288,0.060487998856438525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,10240,0.21381244394514295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,768,0.003724444243643019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,4096,0.25518843862745494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,128,0.015778667396969266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,7168,0.11158222622341579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,32,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,128,0.005533333453867171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,10240,0.014565333724021912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,4096,1.1398746702406142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,7168,0.00944088896115621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,256,0.008189332981904348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,2048,0.04833155539300707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,4096,0.019702222612169053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,16384,0.037969777981440224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,4096,0.00871022211180793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,512,0.012676444318559436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,1024,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,2560,0.013472889032628802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,51200,1.1355466842651367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,5120,0.017496888836224873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,256,0.013960000541475085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,32,0.0028497777465316984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,6144,1.2688168419731987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,4096,0.19068355030483672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,8192,0.17839645014868843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,10240,0.1417777803209093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,3584,0.2998177740308973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,12288,0.01833866702185737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,256,0.0032337777730491427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,6144,0.5957591268751357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,51200,0.32121509975857204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,512,0.007229333122571309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,4096,0.5745671060350206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,12288,0.021329777108298406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,256,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,3584,0.013924444715181986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,128,0.011845333708657159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,512,0.00544088871942626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,6144,0.6069484286838108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,10240,0.31304534276326496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,256,0.03456000155872769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,12288,0.04580355683962504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,256,0.05396355523003472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,7168,0.025948445002237957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,51200,0.21165511343214247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,768,0.0040755557517210645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,5120,0.01718755563100179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,12288,0.19569955931769478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,16384,0.0185217774576611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,512,0.03960799839761522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,2560,0.017467555072572496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,512,0.010799999866220685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,7168,0.4153582255045573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,1024,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,8192,0.013528000149461957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,3072,0.08080799712075128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,51200,0.2706906795501709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,2560,0.0064071110553211635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,64,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,512,0.004855111241340637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,3072,0.10757688681284587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,5120,0.01897333396805657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,768,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,64,0.004259555704063839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,1024,0.011583111352390714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,768,0.007322666545708974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,65536,6.036640167236328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,2048,0.014440889159838358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,3072,0.008209777375062307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,2560,0.12496356169382732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,51200,0.27018221219380695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,12288,0.07849511173036364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,7168,0.10399111111958821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,8192,0.11236000061035156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,128,0.004400889078776042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,65536,0.5748560163709853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,3072,0.027079999446868896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,2048,0.009130666653315226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,6144,0.012925333446926542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,64,0.06054221921496921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,7168,0.016564443707466125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,6144,0.012352888782819113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,1536,0.008919110728634728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,64,0.013583111266295115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,8192,0.00657066661450598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,6144,0.09698488977220322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,3072,0.008548444343937768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,4096,0.16393689314524332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,10240,0.3332933319939507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,3072,0.07850222455130683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,768,0.009889778163697984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,3072,0.25694934527079266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,1024,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,512,0.4285377661387126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,2560,0.09106844663619995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,7168,0.023482667075263128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,64,0.005408888889683618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,32,0.003039999968475766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,1536,0.008654221892356873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,65536,3.1426781548394094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,7168,0.021560000048743352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,4096,0.011927111281288994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,16384,0.4940009117126465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,2048,0.010041777458455827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,51200,3.430074691772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,7168,0.02139555580086178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,32,0.008379555410808986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,6144,0.027469333675172593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,1024,0.011150221857759686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,10240,0.05153510967890421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,4096,1.4715858035617406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,768,0.006626666833957036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,256,0.017471111483044095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,51200,0.05040444599257576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,3072,0.04952266812324524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,5120,0.009296000003814697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,65536,0.13426310486263698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,128,0.004181333300140169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,768,0.23890132374233672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,64,0.16831999354892305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,1024,0.0495431125164032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,1024,0.008470222353935242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,128,0.0030088888274298776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,3072,0.024923556380801733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,7168,0.017994667092959087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,10240,0.11996977859073216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,3584,0.039036442836125694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,2048,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,256,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,4096,0.25333865483601886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,4096,0.08993155426449245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,3584,0.029876444074842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,10240,0.013441777891582914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,256,0.009846222069528367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,51200,0.14002845022413465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,2560,0.012187555432319641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,64,0.0170168893204795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,2048,0.009153777526484596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,2560,0.007270221908887227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,2048,0.029048889875411987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,51200,0.1798479954401652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,12288,0.023656888140572443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,5120,0.02699022160636054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,8192,0.08953511052661473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,2048,0.011162666810883416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,128,0.014128888646761576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,12288,0.4183191193474664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,768,0.013454222016864352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,768,0.00536088893810908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,32,0.0053395554423332214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,12288,0.0506933331489563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,512,0.0061039999127388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,512,0.03356000118785434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,3072,0.05041066805521647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,64,0.007144889069928064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,1024,0.01015644437736935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,512,0.05237955517239041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,16384,7.062287224663629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,51200,0.7204657660590278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,2048,0.011039999624093374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,65536,1.7344658109876845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,256,0.0069066666894488865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,7168,0.05140799946255154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,1024,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,1536,0.08216533395979139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,2560,1.2660409079657662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,64,0.09062400129106309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,8192,0.9007750617133247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,256,0.02164977788925171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,512,0.009002666506502364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,1024,0.08070844411849976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,5120,0.010416000253624385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,3072,0.009834667046864828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,1024,0.004254222330119875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,768,0.012016889121797351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,10240,0.2491999997033013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,8192,0.007557333343558841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,512,0.006406222366624408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,5120,0.025753777888086107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,1024,0.009653333160612319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,1024,0.8527333471510145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,3584,0.023711999257405598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,6144,0.023072888453801472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,65536,0.3818133407168918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,2048,0.050761779149373375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,4096,0.008839999636014303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,1536,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,51200,0.19134399625990126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,1024,0.015467555986510383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,512,0.007879999776681265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,3072,0.006861333217885759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,1024,0.025377778543366328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,32,0.0035786665976047516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,12288,0.02418577836619483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,128,0.004519999855094486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,5120,1.4163519541422527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,12288,0.742686218685574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,64,0.004032000071472592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,12288,0.5421333312988281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,4096,0.010710222025712332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,32,0.006276444428496891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,8192,0.04799377918243408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,51200,1.7274221844143336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,4096,3.820746527777778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,256,0.0074133334888352295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,5120,0.033599111768934466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,3072,0.03300977746645609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,128,0.007763555480374231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,8192,0.04507111178504097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,128,0.1388888888888889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,12288,0.09910933176676433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,7168,0.03346133232116699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,51200,0.9752284155951606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,16384,0.11545599831475152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,16384,0.3482995563083225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,256,0.03345333205329047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,512,0.007304888632562425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,6144,0.024698666400379602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,7168,0.3969279925028483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,2048,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,5120,0.011726222104496427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,1024,0.007040888898902469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,32,0.016978666186332703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,10240,0.07612711191177368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,1536,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,8192,0.021714667479197185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,3072,0.01440711153878106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,3584,0.12178222338358562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,65536,0.9782799614800347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,12288,0.010076444182131026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,3584,0.05699377589755588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,51200,2.487271202935113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,51200,0.6010302437676324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,3072,0.03019733230272929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,12288,0.2843582100338406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,2560,0.014993778533405729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,32,0.0029804443733559716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,768,0.004556444370084339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,4096,0.01368711143732071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,12288,0.06876889202329847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,10240,0.16064711411794028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,10240,2.4431599511040583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,5120,0.018382221460342407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,1024,0.008551999926567078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,7168,0.08048444324069552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,12288,0.06879111131032307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,2560,0.012245333029164208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,1536,0.004752000172932942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,51200,0.34461334016587997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,2048,0.006411555740568373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,64,0.006603555546866522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,12288,0.2776870992448595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,4096,0.012248888611793518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,65536,0.06562399864196777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,51200,0.02500711050298479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,64,0.03995022177696228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,2560,0.039819555150138006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,3072,0.011055111057228513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,3072,0.008237332933478886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,3584,0.008352888955010308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,64,0.04460710949367947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,10240,0.45178312725490993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,2560,0.024111110303137038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,7168,0.017096888687875535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,512,0.0033760000434186724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,5120,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,512,0.007265778051482306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,32,0.019913777709007263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,4096,0.21483200126224092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,4096,0.015987555185953777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,3072,0.1350675556394789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,12288,0.01814399990770552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,12288,0.06789066394170125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,4096,1.1701830757988823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,10240,0.2616746690538195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,768,0.14166755146450466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,3584,0.007246221933099959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,512,0.031231111950344507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,8192,0.023267555567953322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,3072,0.06961511241065131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,512,0.21364355087280273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,256,0.02587466604179806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,1536,0.01221244451072481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,10240,0.2137706677118937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,6144,0.00923822240697013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,3584,0.012308444413873883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,32,0.019655111763212416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,6144,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,256,0.005147555636035072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,768,0.024288889434602525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,256,0.1994862159093221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,10240,0.03514488869243198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,16384,0.06342399782604642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,32,0.0027991111079851785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,2048,0.3772844473520915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,65536,6.545753055148655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,2048,0.036320000886917114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,65536,1.3090470631917317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,512,0.007017778025733099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,256,0.009397333694828881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,65536,0.30319378111097545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,4096,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,1536,0.007368889119890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,6144,0.03270399901602004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,1024,0.05755111243989733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,1024,0.005682666682534748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,768,0.003590222034189436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,8192,0.27496711413065594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,64,0.002963555562827322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,6144,0.5664569007025825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,3584,0.019945777124828763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,32,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,5120,0.013250666360060373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,1024,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,4096,0.17432710859510633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,128,0.005365333209435145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,128,0.03629422187805176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,2048,0.06742578082614474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,65536,0.3694630993737115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,8192,0.13751022020975748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,2048,0.007233777807818518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,2560,0.02288177775012122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,512,0.007866666548781926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,8192,0.04977066649330986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,256,0.007921777665615082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,128,0.03101244568824768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,12288,0.0255422227912479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,768,0.006015111174848344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,5120,0.11081955830256145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,128,0.002946666752298673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,2048,0.008229333493444655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,1536,0.06888888941870795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,64,0.0028355556229750314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,10240,0.006351111249791251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,512,0.13928977648417154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,256,0.0033804445217053094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,3584,0.007428444094128079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,7168,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,32,0.005868444426192178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,4096,0.008242666721343994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,2560,0.03988444474008348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,5120,0.05696977509392632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,768,0.011870221959220039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,64,0.00432977784011099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,1536,0.020849777592553034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,2048,0.004976888911591636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,4096,0.018886221779717337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,128,0.011959111524952782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,4096,0.0814631117714776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,3072,0.01796000036928389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,512,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,8192,0.008393777741326226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,512,0.15230666266547307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,64,0.0028026666906144884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,12288,0.055824889077080615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,5120,0.02070577773782942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,512,0.195177780257331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,768,0.030184000730514526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,128,0.004171555654870139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,5120,0.11383199691772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,7168,0.06137777699364556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,4096,0.08314755227830675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,512,0.007511110769377814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,4096,0.012156444291273752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,10240,0.020941333638297185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,6144,0.035804443889194064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,3584,0.006755555669466655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,8192,0.3699057896931966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,6144,0.38344889216952854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,2048,0.006761777732107375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,768,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,1536,0.06890400250752766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,3584,0.014175110393100314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,7168,0.09515377548005845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,1024,0.00526577772365676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,3072,0.012364444633324942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,128,0.10143022404776679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,1024,0.10125866863462661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,2560,0.012085333466529846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,256,0.0033200000309281876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,3072,0.08814489179187351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,6144,2.3039627075195312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,512,0.012336889074908363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,128,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,128,0.012356444365448423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,7168,0.04563111066818237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,256,0.0063946665161185795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,4096,1.1373626920912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,3584,0.008455110920800103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,256,0.010562666588359408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,12288,0.5122337871127659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,1536,0.0603804455863105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,2048,0.008188444707128737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,768,0.0047048889100551605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,5120,0.09927466842863296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,5120,0.021893334057595994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,3584,0.10346933205922444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,51200,0.8444453345404731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,2048,0.011599111060301462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,32,0.01900533338387807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,3072,0.015644444359673392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,768,0.03290133343802558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,128,0.009683555199040307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,51200,1.1578160391913521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,6144,0.009840000006887648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,64,0.08604088756773208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,4096,0.09774221976598103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,32,0.002741333304180039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,1024,0.007284444239404466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,768,0.01590399940808614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,3072,0.011807111402352652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,3584,0.03732088870472378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,2048,0.006635555376609166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,64,0.038350221183564924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,3072,0.11837955315907796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,1536,0.021691555778185528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,5120,0.00998933364947637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,8192,0.026576000783178542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,10240,0.04331466555595398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,51200,0.1713573270373874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,7168,0.2357635498046875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,128,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,1536,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,51200,0.12448355886671279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,4096,0.015975111060672335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,16384,0.10531377792358398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,768,0.006314666734801398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,16384,0.04704711172315809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,6144,0.8373662100897895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,7168,0.14641422695583767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,10240,0.016632889707883198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,3584,0.14905777242448595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,3584,0.012252444194422828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,7168,0.3018702136145698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,51200,0.9666986465454102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,7168,0.09808533059226142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,768,0.024847999215126038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,3584,0.014698666003015308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,12288,0.34474576844109434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,128,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,4096,0.008050666915045844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,12288,0.036297778288523354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,6144,0.04984088738759359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,1024,0.04606755574544271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,8192,0.9996346367730035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,51200,0.08708710802925958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,2048,0.005422222117582957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,512,0.03590933481852213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,5120,0.07744355334175958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,2048,0.055431110991372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,1536,0.013839999834696451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,65536,0.12574488586849636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,6144,0.018023111753993563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,512,0.09574666950437759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,51200,0.19204178121354845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,3072,0.03344355689154731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,1024,0.012489777472284106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,1536,0.014896889527638754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,256,0.003220444545149803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,4096,0.02628177735540602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,16384,0.030282666285832722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,512,0.01292711082432005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,6144,0.15691999594370523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,8192,0.23684355947706434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,3072,0.021943110558721755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,32,0.007607999775144789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,1024,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,6144,0.14147999551561144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,5120,0.006698666761318843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,128,0.002748444469438659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,10240,0.049682666858037315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,16384,0.5287520090738932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,768,0.07754488786061604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,12288,0.024285333024130926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,7168,3.0760283999972873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,64,0.003620444486538569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,3072,0.013079110946920184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,65536,0.10415289137098525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,512,0.005565333283609814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,128,0.026910222238964502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,1536,0.0335440006521013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,512,0.004499555461936527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,512,0.02416444487041897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,256,0.0053377776510185665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,65536,0.27999377250671387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,768,0.031424889961878456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,10240,0.023071999351183575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,8192,0.5445742077297634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,12288,0.7321919865078397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,7168,0.02843911117977566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,4096,0.011101333631409539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,256,0.00813244448767768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,512,0.04964355627695719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,128,0.022897778285874262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,768,0.03328088919321696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,10240,0.01205244412024816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,1536,0.007272000114123027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,2048,0.213774217499627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,3072,1.7010178036159938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,16384,0.04495199852519565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,65536,0.2401813401116265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,7168,0.019674667053752475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,5120,0.01754755609565311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,768,0.008986666798591614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,8192,0.2811448838975695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,16384,0.04216266671816508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,512,0.01870577699608273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,3584,0.13472621970706516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,1536,0.04513688882191976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,12288,0.05913066864013672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,1536,0.014296000202496847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,128,0.011855999628702799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,12288,0.16205867131551108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,5120,0.01917866700225406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,65536,1.2723076078626845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,1024,0.055916441811455615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,65536,0.016361777981122334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,5120,0.043353776137034096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,4096,0.09805066717995538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,768,0.013807999591032663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,7168,0.026727111803160772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,65536,0.06702666812472872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,32,0.0075502221783002215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,256,0.0040995557275083335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,5120,0.23144798808627656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,5120,0.008950221869680617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,4096,0.027983112467659846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,3584,0.007808888951937358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,128,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,51200,2.9705937703450522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,3584,0.010144888526863521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,256,0.006687110910813014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,3072,0.028232001596026953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,1024,0.04273955689536201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,2560,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,3584,0.022852443986468848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,5120,0.19815644952985975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,16384,0.03352177805370755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,6144,0.13909333282046848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,3584,0.017843556072976854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,8192,0.21260533067915174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,4096,0.011725333001878528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,3584,0.0589688883887397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,10240,3.7437210083007812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,2048,0.007156444092591603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,512,0.01851377718978458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,1024,0.03561866614553664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,512,0.00444444434510337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,8192,0.3587653372022841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,2048,0.005349333501524395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,2560,0.1593386729558309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,1536,0.08377689123153687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,4096,0.008755555583371056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,32,0.003834666477309333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,128,0.004230222354332606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,32,0.03501688771777683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,8192,15.727288140190971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,768,0.07967378033532037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,12288,0.06412533256742689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,512,0.014711999230914645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,512,0.007294221884674496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,8192,0.13674755891164145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,5120,0.042414221498701304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,3584,0.13001510832044813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,3072,0.0476471119456821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,64,0.017055999901559617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,51200,0.22217422061496309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,128,0.030386666456858318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,10240,0.013770666387346057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,5120,0.032560000816980995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,4096,0.008280889027648503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,16384,0.6316133075290257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,51200,0.5761377546522353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,128,0.004935111022657818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,256,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,1024,0.008063111040327284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,6144,0.2136275503370497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,10240,0.03654222355948554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,16384,1.0662497414482963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,8192,0.015267555912335714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,65536,8.197692023383246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,12288,0.026149332523345947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,512,0.003244444520937072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,256,0.0038275557259718576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,1024,0.0047182221379545005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,4096,0.021777777208222285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,64,0.002769777758253945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,8192,0.020848888489935134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,768,0.004571555389298333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,8192,0.12246933248307969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,256,0.005672889037264719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,2560,0.011372444530328115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,64,0.002716444432735443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,128,0.006258666515350342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,3072,0.027991112735536363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,5120,0.011718221836619906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,4096,0.7447111341688367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,65536,0.4385297828250461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,256,0.012287110918098025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,1536,0.035240888595581055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,2048,0.009565333525339762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,1536,0.029652443197038438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,16384,0.032363557153277926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,3072,0.08325955602857801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,64,0.027464000715149775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,8192,0.266461345884535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,16384,0.01296266665061315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,32,0.006700444552633498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,256,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,768,0.35057955318027073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,128,0.0042053332759274375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,768,0.007075555622577667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,12288,0.04929866724544101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,256,0.007426666716734569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,3584,0.018560000591807894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,5120,1.8673163519965277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,8192,0.05897777610354953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,2560,0.008309333688682979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,128,0.014460444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,12288,0.018840889135996502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,65536,0.03081511126624213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,10240,0.05958577659395006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,7168,0.04248355494605171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,768,0.06403111086951362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,128,0.004685333205593957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,64,0.0033644443998734155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,5120,0.016664889123704698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,2048,0.009371555513805812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,8192,0.20762311087714302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,3584,0.014300444059901767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,8192,0.009316444396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,4096,0.07222755750020345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,2048,6.7428800794813375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,768,0.06226488616731432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,768,0.16257777478959826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,32,0.20299466451009116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,12288,0.028930667373869155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,1024,0.006247111078765657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,8192,0.3603795634375678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,256,0.01962488889694214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,10240,0.865939564175076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,65536,0.10654755433400472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,16384,5.402086046006944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,12288,0.011573333707120685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,8192,0.21762400203280977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,64,0.006583999842405319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,2560,0.01826133330663045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,51200,0.38085601064893937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,768,0.1476666662428114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,512,0.0038808890514903595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,65536,0.4285911189185248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,512,0.005665777872006099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,64,0.015441776977645027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,51200,7.762084113226996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,65536,3.9456880357530384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,3584,0.03151199883884854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,7168,0.03464266657829285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,10240,0.16486488448248968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,768,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,65536,1.800524393717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,12288,0.028311110205120508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,51200,0.3598053455352783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,65536,0.013295110729005603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,5120,0.05671911107169258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,10240,0.09975910849041408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,6144,0.01129155523247189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,65536,0.10959821939468384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,16384,0.05208355519506666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,7168,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,3584,0.006741333338949416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,1024,0.02164711058139801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,64,0.0028826666788922418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,65536,0.21720355086856416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,32,0.03195910983615451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,10240,8.58501942952474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,1536,0.00999644481473499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,1024,0.006364444477690591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,51200,0.10014666451348199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,1024,0.006797333144479328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,512,0.008817777865462834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,6144,0.012220444778601328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,3584,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,3072,0.008091555701361762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,768,0.012532444463835822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,128,0.0058746664888328975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,128,0.008232889076073965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,128,0.009345778160625035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,7168,0.02051822178893619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,768,0.003680888977315691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,512,0.03544355432192484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,2560,0.11057511303159927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,7168,0.07315288649664985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,4096,0.006166222194830577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,65536,0.03902133305867513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,1536,0.004555555681387584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,10240,0.009392889009581672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,768,0.007424888511498769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,8192,0.1845724450217353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,512,0.009619555539555019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,128,0.0351235568523407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,3072,0.11041066381666396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,1024,0.014715555641386243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,7168,0.012701333396964602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,3072,0.01497688889503479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,10240,0.042648888296551175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,5120,0.49919912550184464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,65536,0.4759048885769314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,1024,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,3584,0.01962488889694214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,10240,0.046894222497940063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,10240,6.182741377088759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,51200,0.7175768746270074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,8192,0.41931290096706814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,1024,0.017893332574102614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,32,0.0038862224254343244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,6144,0.019124444988038804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,256,0.0031519999934567344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,512,0.16762133439381918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,5120,0.24564266204833984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,7168,0.4993297788831923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,10240,0.22624177402920195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,768,0.010129777921570672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,4096,0.01223288890388277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,1536,0.058055996894836426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,8192,0.02994933393266466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,2048,0.16246133380466038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,8192,0.01440533333354526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,5120,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,256,0.02195377813445197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,2048,0.029751112063725788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,1536,0.005728000154097875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,16384,0.018990221950742934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,6144,0.023950222465727065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,10240,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,10240,0.0857342216703627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,128,0.004423999951945411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,16384,0.12103288703494602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,10240,0.03645955522855123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,3584,3.6707411872016062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,256,0.008766222331258986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,32,0.003335111050142182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,51200,0.017099555995729234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,768,0.005058666484223472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,3584,0.015763555963834126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,256,0.00544177782204416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,768,0.08875733613967896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,8192,0.015216888652907478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,768,0.003738666574160258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,1024,0.0041475556790828705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,128,0.010185778141021729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,4096,0.027661333481470745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,8192,0.10179199775060017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,3072,0.17588444550832114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,3072,0.0336497790283627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,7168,0.014763555592960782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,3072,0.20706221792432997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,7168,0.027519999278916255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,3072,0.11199466387430827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,768,0.05688533518049452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,5120,0.19656533665127227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,32,0.030046221282747056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,10240,0.01551555593808492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,3584,0.0797262191772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,16384,0.534130679236518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,256,0.007439110842016008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,3072,0.582557307349311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,51200,0.33125511805216473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,256,0.0037875556283526947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,12288,0.025628444221284654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,5120,0.013159111142158508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,1536,0.04997600118319193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,3584,0.05852622456020779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,10240,0.020215999748971727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,3072,0.013424889081054263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,5120,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,51200,0.07482488950093587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,256,0.00462400002612008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,256,0.003987555702527364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,2560,0.009362666971153682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,768,0.004096888833575779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,128,0.0046977777447965415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,32,0.006693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,51200,0.2692728837331136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,16384,0.06124266650941637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,12288,3.407115512424045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,3584,0.008785777621799046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,3584,0.019754666421148512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,2560,0.02258577777279748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,51200,20.595104641384548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,2560,0.018575110369258456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,51200,0.05345244540108574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,512,0.03372977839575873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,8192,0.07649066713121203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,512,0.06885421938366361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,10240,3.519737879435221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,256,0.01959022217326694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,8192,0.0158915552828047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,8192,0.15648533238304987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,65536,0.9072888692220052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,128,0.04458399944835239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,8192,0.038395557138654925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,128,0.003091555502679613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,768,0.008182221816645728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,32,0.01093155559566286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,2048,0.007010666860474481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,2560,0.02364088926050398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,12288,0.02749155627356635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,512,0.013990221752060784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,65536,0.21925066577063668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,768,0.00498577786816491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,16384,0.8509768909878201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,4096,0.0897902250289917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,1536,0.3186417685614692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,128,0.01921244462331136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,4096,0.08762488762537639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,2560,0.013531555732091268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,7168,0.012708444562223224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,512,0.007675555845101674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,2560,0.04602844516436259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,3584,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,16384,0.07533866829342313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,1536,0.010847110715177325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,768,0.014904888139830695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,3584,0.058408002058664955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,5120,0.03562488820817735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,512,0.007658667034573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,4096,0.018820444742838543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,2560,0.0339955555068122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,51200,0.0603297750155131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,6144,0.047467556264665395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,2048,0.009566222627957663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,10240,0.044799112611346774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,7168,0.12606578403049046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,4096,0.01368977791733212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,512,0.0035991109907627106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,5120,0.010632000035709806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,1536,0.007179555793603261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,3072,0.1310675541559855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,768,0.05428355601098803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,3072,0.008481777376598781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,768,0.09141066339280869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,64,1.4286844465467665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,5120,0.21717600027720133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,5120,0.023337778117921617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,16384,0.2821999920739068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,2048,0.016018667154841952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,16384,0.038189331690470375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,3584,0.0092613332801395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,65536,0.2107635604010688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,65536,11.314242892795138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,4096,0.07156799899207221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,16384,0.0335822237862481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,7168,0.010849778023031024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,768,0.18131200472513834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,128,0.005234666582610872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,51200,0.099181334177653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,2560,0.013605333036846586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,1024,0.019923556182119582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,8192,0.015288889408111572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,2048,0.09742133484946357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,12288,1.1656808853149414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,7168,0.7951599756876627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,51200,0.2998880015479194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,8192,0.015254222684436373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,3584,0.0573822193675571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,2560,0.015223999818166098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,51200,0.09514311287138198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,5120,0.012258666257063547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,12288,8.185226440429688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,16384,0.04528177777926127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,768,0.004419555680619346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,5120,0.1379342211617364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,1536,0.009532444179058075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,51200,0.6643466419643825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,1536,0.010592888626787396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,512,0.003875555677546395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,4096,0.23413332303365073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,10240,0.01744622157679664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,32,0.007435555259386699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,5120,0.010176888770527309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,1024,0.0070879997478591064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,5120,0.12816088729434544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,51200,0.1985466612709893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,4096,0.0787208875020345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,2048,0.027497778336207073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,2048,0.28990311092800564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,10240,0.1644622219933404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,4096,0.010332444475756751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,1024,0.03587999939918518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,51200,5.622546725802952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,2048,0.05163377854559156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,10240,0.0123511114054256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,16384,0.408876445558336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,1024,0.013352000051074557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,64,0.008591999610265097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,3584,0.014128888646761576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,65536,0.2843048837449816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,3072,0.08384355571534897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,256,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,65536,0.2703804439968533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,16384,0.17648533980051676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,1024,0.02664622167746226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,128,0.015863110621770222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,768,0.006627555522653792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,12288,0.1805493301815457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,1536,0.025802666942278545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,2048,0.01479555500878228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,5120,0.1480284399456448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,7168,0.013174222575293647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,2560,0.0918053322368198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,64,0.10537511110305786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,65536,0.34249067306518555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,3584,0.007702222300900354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,51200,0.2105564408832126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,1024,0.2007991075515747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,51200,0.43318843841552734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,64,0.2448106606801351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,1024,0.07025333245595296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,10240,0.17120444774627686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,65536,0.3237653308444553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,16384,0.019304000669055514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,2048,0.02427822185887231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,768,0.004555555681387584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,1024,0.029905776182810467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,768,0.13151644335852727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,8192,0.1419102218416002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,3584,0.029613332615958318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,512,0.005249777601824866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,65536,0.06735377841525607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,3072,0.011109333071443768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,1024,0.1964177820417616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,12288,0.4037155575222439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,1536,0.02627022233274248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,6144,0.031106667386160955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,5120,0.045485334263907544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,32,0.002985777747299936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,512,0.0037617778612507712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,128,0.020062221421135795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,3584,0.009857777920034196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,1024,0.03672000103526645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,768,0.005110222018427319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,51200,0.0603991084628635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,4096,0.01774311065673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,51200,0.30299022462632924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,2048,0.062126219272613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,3584,0.009429333110650381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,256,0.12007289462619358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,768,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,3072,0.12207199467553033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,2048,0.03212889035542806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,1024,0.012654222548007965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,10240,0.5882524384392632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,256,0.018152889278199937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,8192,0.15187733703189427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,51200,0.18101867039998373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,1024,0.010288000106811523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,8192,0.07480622000164457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,51200,0.059761775864495166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,2560,0.008681777450773451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,10240,0.006129777679840724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,10240,0.057942224873436816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,12288,0.05189422104093763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,1024,0.005676444619894028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,7168,0.03350844317012363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,512,0.005402666827042897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,7168,0.06747377581066556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,5120,0.02540622154871623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,7168,0.03038933210902744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,7168,0.01642577846844991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,2048,0.008151999778217739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,256,0.00656355544924736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,1536,0.00832088871134652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,768,0.026836444934209187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,1536,0.021398221453030903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,512,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,1024,0.004280888785918553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,512,0.00852711084816191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,128,0.00508088908261723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,768,0.005706666658322017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,1024,0.008709333009190029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,256,0.00360622215602133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,768,0.005751111027267244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,1536,0.012244444754388599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,3072,0.007649777664078607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,1024,0.00397777764333619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,512,0.0053839998112784485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,16384,0.13663466771443686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,5120,0.03174933460023668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,3584,0.10210222005844116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,2048,0.09353422456317478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,6144,0.20757688416375053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,1024,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,256,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,51200,1.7341652976142035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,5120,3.6415350172254772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,5120,0.10824622048272027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,768,0.012721777790122561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,64,0.26107467545403373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,2048,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,2560,0.010072000324726105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,16384,0.020942222740915086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,6144,0.09566489193174575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,256,0.006791111081838608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,16384,0.8447982470194498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,1024,0.00406577769252989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,32,0.01979111135005951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,2560,0.014108444253603617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,2048,0.3927466604444716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,65536,0.2931973404354519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,12288,4.571524302164714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,128,0.033492445945739746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,7168,0.04355022311210632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,7168,0.18339022000630698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,16384,0.1986293395360311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,128,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,256,0.007280000381999546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,64,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,3584,0.00887288898229599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,12288,2.576037300957574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,768,0.02238133384121789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,51200,0.04969955484072367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,6144,0.08013511366314359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,8192,0.015669332610236276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,6144,0.10488888952467178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,10240,0.03512710995144314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,1024,0.0064497776329517365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,32,0.04876800047026741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,6144,0.1810471084382799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,512,0.005583999885453119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,256,0.007169778148333232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,2048,0.013714666995737286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,4096,0.0184888889392217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,51200,0.06249155600865682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,51200,3.1109049055311413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,2048,0.03534222311443753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,64,0.002965333354141977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,512,0.019620445039537217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,2048,0.058209776878356934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,8192,0.46326933966742617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,64,0.005658666706747479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,768,0.11416355768839519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,51200,0.8649653328789605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,16384,0.0480968885951572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,65536,33.22357516818577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,2048,0.04193422198295593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,3584,0.03080355458789402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,16384,0.10503022538291083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,6144,0.010135110881593492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,64,0.00657066661450598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,4096,0.12406310770246719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,3584,0.006746666712893381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,12288,0.21374311712053087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,2048,0.0443520016140408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,4096,0.13612355126274958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,4096,0.01477688882086012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,512,0.04125866625044081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,3584,0.22322132852342394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,3584,0.008029333419269985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,1024,0.01497333414024777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,3584,0.05916533205244276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,3072,0.009644444617960189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,3072,0.014515555567211576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,4096,0.5709724426269531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,512,0.16559733284844294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,12288,1.3605368932088215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,65536,1.8161048889160156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,64,0.007232888705200619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,512,0.005382222019963794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,5120,0.46415466732449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,64,0.0028880000528362063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,2560,0.007779555188284979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,8192,0.05385422375467089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,3072,0.045927110645506114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,8192,0.04581244455443489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,4096,0.03698044353061252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,512,0.004013333469629288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,768,0.21262666914198133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,1024,0.009146666361225976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,64,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,256,0.004363555461168289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,51200,13.590065002441406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,1024,0.05520888831880358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,2560,0.009841777384281158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,1536,0.060481780105166964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,12288,0.06272977590560913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,7168,0.019292443990707397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,6144,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,10240,0.031445334355036415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,6144,0.03804533349143134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,32,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,256,0.0031102223114834893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,3072,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,12288,0.055480887492497764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,2560,0.012114666402339935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,768,0.004591111093759537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,128,0.0028080000645584534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,4096,0.053857776853773326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,5120,0.008187555604510838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,10240,0.0358586675590939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,2048,0.016356445021099515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,4096,0.01200177768866221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,12288,0.8305235438876681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,5120,0.012466666599114736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,2560,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,51200,0.12082578076256646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,51200,1.5313680436876085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,5120,0.021168889270888433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,8192,0.029044446018007066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,1536,0.006225777582989798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,2560,0.03135644396146139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,1536,0.004671999977694618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,1536,0.02230666743384467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,3072,0.06419999731911553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,16384,0.023524444964196947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,128,0.015537778536478678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,128,0.004658666749795278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,32,0.002516444358560774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,10240,0.19573599762386748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,2560,0.009691555466916826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,1024,0.04768800073199802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,256,0.02086399992307027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,1536,0.009849777652157677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,512,0.005586666779385672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,2560,0.025957332717047796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,64,0.0053226666318045715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,1024,0.028510222832361858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,64,0.020597333709398907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,3072,0.011483555866612328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,3584,1.1786862479315865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,128,0.009771555662155151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,3584,0.018200000127156574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,3584,0.05302933189604017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,12288,0.0181013329161538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,32,0.0030871110243929755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,256,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,1024,0.019371555911170114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,51200,0.0542719993326399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,128,0.006565333240562015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,1024,0.013518222504191928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,768,0.003670222229427762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,2560,0.22661778661939833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,3072,0.08512622117996216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,1536,0.006591111007663939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,128,0.03313600023587545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,512,0.00905777762333552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,2048,0.031559109687805176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,8192,0.011496889094511667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,3072,0.23450401094224718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,64,0.01672266589270698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,2048,0.21357599894205728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,3584,0.022504000200165644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,2560,0.008147555920812819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,128,0.0034942221310403612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,1536,0.112972444958157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,128,0.003301333429084884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,5120,0.0825022194120619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,51200,0.0827253328429328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,2048,0.01581333412064446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,64,0.00535999983549118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,512,0.005683555371231503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,256,0.04974933465321859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,512,0.0031866667171319327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,768,0.004403555558787452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,5120,0.032709333631727434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,65536,0.07016177972157796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,7168,0.027229333917299908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,32,0.013037333057986366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,65536,0.6295982466803657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,1536,0.022848000129063923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,6144,0.5716604656643337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,32,0.004267555558019214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,10240,0.047120889027913414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,256,0.009565333525339762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,12288,0.014647111296653748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,16384,0.12045866913265652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,65536,1.2983457777235243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,16384,0.10447555780410767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,64,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,2048,0.012216889195972018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,128,0.006236444330877728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,1536,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,65536,0.09358488851123387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,512,0.29010311762491864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,4096,0.1413173278172811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,51200,0.058470222685072154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,12288,0.09550222423341538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,768,0.012839111189047495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,65536,0.2742408911387126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,512,0.01977688901954227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,65536,0.07489066653781466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,7168,0.10821600092781915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,12288,0.09088089068730672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,2560,0.04603911108440823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,51200,0.3842008908589681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,8192,0.015669332610236276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,16384,1.6450284322102864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,16384,1.363960901896159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,5120,0.1521377828386095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,65536,0.22542754809061685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,8192,0.04293333490689596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,10240,0.06184799803627861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,51200,0.08171733220418294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,8192,0.016198222835858662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,2048,0.04904799991183811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,65536,0.30032621489630806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,64,0.010475555227862464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,5120,0.01623911162217458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,4096,0.14208000236087376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,7168,0.020429333051045735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,1536,0.004639111045334074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,3072,0.0075626663035816616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,51200,0.6245111359490289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,512,0.0036755556033717263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,4096,0.014922666880819531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,12288,0.8178266419304742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,51200,0.09245866537094116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,2048,0.006739555547634761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,2560,0.014688889185587565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,512,0.21524533960554335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,768,0.04236977630191379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,3584,0.03453866640726725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,6144,0.0541928874121772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,1024,0.461043569776747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,512,0.005743999862008625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,128,0.00282577777074443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,12288,0.05113777849409315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,3072,0.08598311079872979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,32,0.002850666642189026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,2560,0.21147199471791586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,51200,0.05535911189185249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,256,0.012897777888509961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,32,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,6144,0.09259733226564194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,512,0.0399982233842214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,16384,0.1274053388171726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,51200,4.0339029100206165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,512,0.0047973332305749255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,768,0.12217156092325847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,3072,0.08893333541022407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,128,0.004870222260554631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,256,0.024506666594081458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,6144,0.11580444706810845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,256,0.008624888956546783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,2560,0.02904711167017619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,1024,0.01034755590889189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,4096,0.06920799944135878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,8192,0.0892533328798082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,1024,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,4096,0.008652444514963362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,3072,0.020049777295854356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,3584,0.006166222194830577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,10240,0.016875555117925007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,2048,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,5120,0.07311199771033393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,1536,0.018740443719757926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,7168,0.07602844635645549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,6144,0.03158400124973721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,6144,0.16842310958438447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,3072,0.40282310379876035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,64,0.005487111293607288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,768,0.04031022058592902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,256,0.009628444082207149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,5120,0.015692444311247934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,3584,0.008106666306654612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,2048,0.0351511107550727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,10240,0.019136889113320243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,12288,0.012089778151777057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,2560,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,6144,0.01110222190618515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,65536,0.088664001888699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,1536,0.005588444570700328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,6144,0.01665599975321028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,16384,0.06902755631340875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,3072,0.012523555921183692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,3072,0.008538666698667739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,5120,0.019337778290112812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,32,0.002712888850106133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,12288,0.04886044396294487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,3072,0.009771555662155151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,768,0.05043288734224108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,6144,0.006755555669466655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,5120,0.0558533337381151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,64,0.002692444456948174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,64,0.003779555360476176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,10240,0.25033243497212726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,64,0.008372444245550368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,12288,0.024135111106766596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,2560,0.08954666720496283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,128,0.007010666860474481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,65536,0.9607644610934787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,256,0.005104889058404499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,128,0.019853333632151287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,5120,0.09052799807654487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,5120,0.06389066908094618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,768,0.012840000291665396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,7168,1.3093013763427734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,51200,0.5424817932976617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,768,0.009328888522254096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,8192,0.03472000029351976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,4096,0.014548444085650973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,65536,0.2773439884185791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,768,0.00582755563987626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,8192,0.012705778082211813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,12288,0.06936622328228421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,16384,0.05637066894107395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,51200,0.07162755727767944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,3072,0.12072532706790501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,5120,0.012200000385443369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,128,0.004285333471165763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,16384,0.026037333740128413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,8192,0.030793779426150854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,7168,0.016034666034910414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,1536,0.021203555994563635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,768,0.030893332428402368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,5120,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,256,0.013474666410022311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,128,0.004914666629499859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,8192,0.8379831314086914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,64,0.009465778039561378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,1536,0.00903200027015474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,5120,0.012400888734393649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,2560,0.5979484452141656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,512,0.013241777817408243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,3584,0.16812533802456328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,64,0.006106666806671355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,12288,0.02176355487770504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,2048,0.005920889063013925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,5120,0.06359110938178168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,4096,0.010202666951550378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,256,0.014096000128322177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,65536,7.935776604546441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,768,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,16384,0.5000435511271158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,2560,0.008352888955010308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,65536,0.1205564472410414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,32,0.0036551112102137674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,1536,0.012423111332787408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,3072,0.02297244469324748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,12288,0.020420443680551317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,512,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,64,0.051408890220854014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,7168,0.031585776143603854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,6144,0.19920621977912056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,2048,0.006979555719428592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,2048,0.01366311146153344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,512,0.003656889001528422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,6144,0.025624000363879736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,7168,0.07198844353357951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,12288,0.19132178359561494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,12288,0.03130044539769491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,2048,0.007780444290902879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,512,0.005988444305128521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,3072,2.165970696343316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,8192,0.061807109249962695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,1024,0.04695111182000902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,2560,0.009145778086450364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,51200,4.682860480414496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,7168,0.30404888259039986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,3584,0.011026666396194033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,51200,1.1799937354193795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,1536,0.0074240002367231585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,768,0.0657235582669576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,7168,0.09167999691433376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,5120,0.6645493507385254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,1536,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,512,0.013068444199032254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,8192,0.19822133911980522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,768,0.02392622166209751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,6144,0.02773244513405694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,64,0.0058746664888328975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,16384,0.049888001547919385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,512,0.005568888866239124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,1536,0.004927111168702443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,512,0.008919999831252629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,3072,0.04758755697144402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,51200,5.737054189046224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,2048,0.007738666401969061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,2048,0.006032889087994893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,32,0.003714666598372989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,128,0.0031271111220121384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,3072,0.9685280058119031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,51200,0.548692438337538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,8192,0.027453333139419556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,512,0.06991644700368245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,2560,0.016844444804721408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,8192,0.03820977939499749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,1536,0.007400888535711501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,3584,0.014345778359307183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,51200,0.1699742211235894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,256,0.01665511065059238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,3584,0.257076448864407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,4096,0.05860000186496311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,12288,0.2800515492757161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,3584,0.008591999610265097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,10240,0.013591110706329346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,12288,0.015075556106037565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,128,0.003742222156789568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,3072,0.00756888919406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,256,0.005671111245950063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,7168,0.014053333136770459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,12288,0.6520142025417751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,5120,0.48650312423706055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,128,0.003318222239613533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,256,0.14260355631510416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,64,0.003967111309369405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,2560,0.01460088955031501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,6144,0.0502106679810418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,5120,0.032130665249294706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,2560,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,12288,0.0909208920266893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,512,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,512,0.032815112007988825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,768,0.0317520002524058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,64,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,32,0.0028693332440323303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,1536,0.04741600155830383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,7168,0.15427199999491373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,1536,0.13287199868096247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,4096,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,65536,0.4320640034145779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,512,0.024352888266245525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,3072,0.022016000416543748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,3584,0.022690667046440974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,768,0.0064026667839950984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,256,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,2560,0.0070017774899800616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,3072,0.021935110290845234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,256,0.005049777941571342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,5120,0.2561146683163113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,3072,0.009822222093741098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,5120,0.06866577598783705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,8192,0.01570933394961887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,3072,0.07908355527453952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,3584,0.3263173368242052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,64,0.005421333428886201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,65536,0.009728888670603434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,128,0.0045422220395671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,12288,0.017120889491505094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,768,0.01844888925552368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,16384,8.013809204101562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,6144,0.012240889171759287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,8192,0.17981689506106904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,1536,0.03851200143496195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,51200,9.959767659505209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,16384,0.03980355461438497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,5120,0.10486755768458049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,65536,0.5164240201314291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,256,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,256,0.007877333296669854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,1024,0.08593244685067071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,16384,0.0306275553173489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,3584,0.03874399926927354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,2560,0.007703111403518253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,7168,0.5285777515835232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,12288,0.8764044443766276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,65536,0.1642035510804918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,256,0.022872888379626807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,65536,3.397392908732096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,6144,0.023936000135209825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,3584,0.8434506522284614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,7168,0.012879110872745514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,64,0.002806222273243798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,65536,0.20213777489132354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,12288,0.044459554884168834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,512,0.008401778009202745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,1536,0.13028444184197321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,51200,0.12538933753967285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,6144,0.253528012169732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,1536,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,16384,0.2663697666592068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,128,0.007869333028793335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,3072,0.07673599984910753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,7168,0.3868417739868164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,10240,0.040375110175874494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,2560,0.04721422327889336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,256,0.005184000151024925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,768,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,8192,0.029862221744325426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,3584,0.01315822203954061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,8192,0.05328266819318136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,6144,0.19987644089592826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,2048,2.5335138108995223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,1536,0.015412444869677225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,128,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,4096,0.011136000355084738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,6144,0.015191111299726697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,1024,0.016515556308958266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,7168,0.0793271130985684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,2048,0.01293777757220798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,1536,0.006785777707894643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,768,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,1536,0.007027555671003129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,16384,0.10800355672836304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,2560,0.03387555480003357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,2048,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,8192,0.5086480246649848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,6144,0.09182222021950616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,4096,0.01992177797688378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,1024,0.01257066677014033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,64,0.007289778027269576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,1024,0.004483555754025777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,7168,0.007709333466158972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,51200,0.41873955726623535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,4096,0.009248889154858058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,1024,0.011580444044537015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,128,0.036792887581719294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,32,0.007151111132568783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,2560,0.010145777629481422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,4096,0.010386666489972008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,51200,0.980778694152832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,7168,2.052264955308702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,4096,0.011403555671374003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,6144,0.042958223157458834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,12288,0.028636445601781208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,4096,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,4096,0.015152888165579902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,3072,0.00980444418059455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,512,0.004265777766704559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,512,0.006364444477690591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,3072,0.021353777911927965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,512,0.07228533426920573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,10240,0.07115466727150811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,1536,0.03966222206751505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,5120,0.09398933251698811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,64,0.0053324442770746015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,12288,0.037821332613627114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,16384,0.22609512011210123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,512,0.02743911080890232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,32,0.002647111192345619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,8192,0.02201066745652093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,7168,0.03398311138153076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,7168,0.020223110914230347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,51200,0.08177955283059014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,65536,0.116321775648329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,4096,2.076523462931315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,128,0.02351644469632043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,768,0.01016088906261656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,1536,0.20383911662631563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,12288,4.449248843722873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,128,0.012303999728626676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,3584,0.1896444426642524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,16384,0.04863644308514065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,32,0.0029102222373088202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,65536,0.3868310981326633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,256,0.005139555368158552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,10240,0.4282906585269504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,3072,0.026712000370025635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,16384,0.023248889380031165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,64,0.008848889006508721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,8192,0.04757510953479343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,32,0.006477333191368315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,7168,0.05173422230614556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,51200,0.12106755044725205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,5120,0.010240000155236986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,768,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,5120,0.01202755586968528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,6144,0.012597333225939008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,5120,0.028417779339684382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,65536,0.29352977540757924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,16384,0.04442044430308872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,8192,0.19708177778455946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,3584,12.366053263346354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,5120,0.013482666677898832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,2560,0.03231822119818793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,128,0.006559999866618051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,64,0.00254133323000537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,512,0.025847999585999384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,64,0.002959111084540685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,8192,0.021643555826610986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,6144,0.07708355453279284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,32,0.002965333354141977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,512,0.004960888789759742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,8192,1.8455564710828993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,65536,0.09531644317838882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,128,0.00629688882165485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,5120,0.35716621081034344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,3072,0.022934221559100684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,12288,0.024159999357329473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,5120,0.021054221524132624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,4096,0.006301333506902059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,16384,0.01904088921017117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,6144,0.03231288989384969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,12288,0.04523288872506884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,32,0.09750577476289536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,64,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,2048,0.05977777640024821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,8192,0.039692444933785334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,64,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,4096,0.006637333167923822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,8192,0.011607999602953592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,6144,0.011830222275522021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,5120,0.006583111153708563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,51200,0.014481777946154276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,1024,0.011892444557613797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,16384,0.0572648909356859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,5120,0.008625778059164682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,1536,0.007111999723646376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,64,0.002489777695801523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,3072,0.038197331958346896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,10240,0.032948444286982216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,16384,53.59356011284722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,10240,0.021568000316619873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,51200,10.121331956651476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,1536,0.008573333422342936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,1024,0.007592000067234039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,3584,0.02346755564212799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,1024,0.006948444578382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,16384,0.023316444622145757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,16384,0.15483821762932673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,256,0.01998844411638048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,5120,0.013088888592190213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,64,0.005240889059172736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,5120,0.011335999601417117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,128,0.010610666539933948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,512,0.00851288851764467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,4096,0.008402666283978356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,10240,0.02886755598915948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,128,0.004416000097990036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,1536,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,4096,0.11105244689517552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,2560,0.014189332723617554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,12288,0.08387022548251682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,3072,0.010780444575680627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,3584,0.15144444836510554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,1024,0.02054577734735277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,3072,0.009166222479608323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,3584,0.007288888924651676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,12288,0.07687111033333673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,256,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,65536,0.12683644559648302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,2048,0.02593244446648492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,1024,0.023176888624827068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,6144,0.008762666748629676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,7168,0.018067555295096505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,256,0.004544888933499654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,51200,2.532449722290039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,3072,0.04174222217665779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,512,0.004643555730581284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,2560,0.09009689092636108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,1536,0.013320888910028668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,5120,1.3096870846218533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,7168,0.017109332813156974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,256,0.003915555361244413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,16384,7.448337131076389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,3072,0.12035822868347168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,10240,0.01922044489118788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,1024,0.007794666621420119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,768,0.018600889378123812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,1536,0.015814221567577787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,128,0.02035555574629042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,4096,0.008023111356629265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,16384,0.07644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,65536,0.03068000078201294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,6144,0.03978666663169861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,2560,0.016061334146393668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,10240,0.04394222299257914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,5120,0.014713777436150445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,51200,0.046576887369155884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,65536,0.4282248814900716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,3072,0.036537776390711464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,768,0.006151111175616582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,12288,0.18132355478074816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,3072,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,256,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,1536,0.04253422220547994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,64,0.04641155401865641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,3584,0.014368888404634265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,16384,0.042731556627485484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,8192,0.20066932837168375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,10240,0.23288710912068686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,10240,0.12475021680196126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,12288,0.47804800669352215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,3584,0.2827582094404432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,4096,0.11653244495391846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,4096,0.12162044313218857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,4096,0.2690471013387044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,5120,0.020288000504175823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,7168,0.27366844813028973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,5120,0.01148888882663515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,64,0.006245333287451003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,7168,0.010245333115259806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,1536,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,1024,0.00721955547730128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,32,0.0028080000645584534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,3072,0.01686044368478987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,32,0.005144888742102517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,32,0.0030533334033356774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,10240,0.13328533702426487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,1536,0.009986666341622671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,12288,0.019762666689025033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,12288,0.04194666610823738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,3584,0.39825956026713055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,16384,0.05562844541337755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,2560,0.01368444412946701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,1024,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,65536,1.7373938030666773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,2048,0.09877689017189874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,16384,0.025292444560262892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,1536,0.015045333239767285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,1024,0.058187557591332324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,2560,0.08024355438020495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,65536,0.1059911118613349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,64,0.005345777918895085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,2048,0.011905777785513135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,7168,0.02787999974356757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,2560,0.014491554763582019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,768,0.024640889631377325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,32,0.005316444569163852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,3584,0.019357333580652874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,8192,0.03693777653906081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,256,0.009266667068004608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,4096,0.012474666866991254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,4096,0.07618221971723768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,128,0.40673332744174534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,128,0.013769778112570444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,51200,0.06239022148980034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,512,0.009961778091059791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,4096,0.015463110473420886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,64,0.11912622716691758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,256,0.0046951112647851305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,1536,0.011390222443474663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,10240,0.36753778987460667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,12288,0.020440000626775954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,12288,0.18131289217207167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,2560,0.634319093492296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,10240,0.008018666671382057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,10240,0.0417902237839169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,16384,0.37029867702060276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,6144,0.24831822183397081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,3072,0.07383288939793904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,512,0.005412444472312927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,128,0.18507822354634604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,1536,0.04898311032189263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,1024,0.004413333204057482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,6144,0.11306577258639866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,51200,0.4688426653544108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,51200,22.897581312391495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,4096,0.07926044199201795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,10240,0.16161689493391249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,768,0.005664000080691443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,12288,0.009953777823183272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,512,0.03480089041921828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,1536,0.019542222221692402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,1024,0.019691555036438834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,12288,0.1568088928858439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,8192,0.014211555322011312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,12288,0.015683554940753512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,5120,0.017854221993022494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,10240,0.754217783610026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,5120,0.13013777467939588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,8192,0.06064889166090223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,16384,0.04227822356753879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,1536,0.034407999780442976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,7168,0.011853333148691388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,6144,0.025944888591766357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,512,0.015951111912727356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,2560,0.05086044470469157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,512,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,256,0.06748888889948527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,10240,0.39793067508273655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,5120,0.31271200709872776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,12288,0.042176889048682324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,10240,0.01870755520131853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,65536,1.2330008612738717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,2560,0.8024097548590766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,51200,0.6105662451850044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,65536,0.4613964292738173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,65536,0.13972532749176025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,512,0.06282222270965576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,8192,0.06168888674841987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,256,0.011515555282433828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,10240,0.13768800099690756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,8192,0.02423200011253357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,64,0.007073777417341868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,32,0.0026622222115596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,2560,0.01666933298110962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,8192,0.0641626649432712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,6144,0.010311110979980892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,51200,0.8255022366841634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,3072,0.01644711030854119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,2048,0.008755555583371056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,4096,0.016509332590632968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,4096,0.3108355469173855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,3584,0.017504000001483493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,3072,0.05890488624572754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,12288,0.04239111145337423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,5120,0.01905155513021681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,8192,0.26975732379489475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,4096,0.09376444419225057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,51200,0.4770977762010362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,51200,0.38003910912407773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,128,0.005195555587609609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,32,0.007575111256705389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,4096,0.07017244233025445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,8192,0.02797244323624505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,256,0.005402666827042897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,51200,0.11363377836015488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,6144,0.08223199844360352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,2048,0.0945075551668803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,64,0.005507555686765247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,3072,0.0064239998658498125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,12288,2.497830284966363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,768,0.004767111192146937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,2560,0.009311111436949836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,128,0.003891555385457145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,7168,0.01459200017982059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,7168,0.23713776800367567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,4096,0.11427377329932319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,768,0.015294222368134392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,5120,0.028468443287743464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,256,0.0052791109515561005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,8192,0.4493858019510905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,16384,0.5322391192118326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,3072,0.007304888632562425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,3072,0.010414222048388587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,1536,0.010187555518415239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,10240,0.010841777755154504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,512,0.015100444356600443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,4096,0.009955555200576782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,6144,0.09272266758812799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,65536,2.2221546173095703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,12288,5.475582970513238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,5120,0.027302222119437322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,1024,0.005574222240183089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,128,0.01829688913292355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,16384,0.027444443768925134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,3072,0.009152889251708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,7168,3.4154027303059897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,2048,0.06344444221920438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,16384,0.013076444466908773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,2560,0.014493332968817817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,768,0.0184871107339859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,1024,0.01258399999803967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,768,0.016727111405796476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,128,0.003121777748068174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,128,0.015291555060280694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,3584,1.9806836446126301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,1536,0.1135057740741306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,8192,0.045520888434516064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,128,0.014914666612943014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,2560,0.03110044532352024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,512,0.01330311099688212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,768,0.5836906433105469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,10240,5.628341250949436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,1536,0.06315110789404975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,7168,0.5922800170050727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,1024,0.042621334393819175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,512,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,64,0.013019555144839816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,8192,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,8192,0.04654133319854736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,65536,4.116536034478082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,12288,0.08008355564541288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,65536,0.11372533109452988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,2048,0.006530666516886816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,4096,0.02052177819940779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,128,0.02201066745652093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,10240,0.1815191109975179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,32,0.005136000199450387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,3072,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,128,0.005143110950787862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,12288,0.8177626927693685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,1536,0.010016000105275048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,512,0.005056000004212062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,1024,0.004878222114510006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,768,2.5213911268446183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,768,0.009472000102202097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,5120,0.028816001282797918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,256,0.003383999897374047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,768,0.006839111033413146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,768,0.08905333280563354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,2048,0.016746666696336534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,256,0.0199635558658176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,64,0.00305600009030766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,3584,0.2189617819256253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,256,0.012120000190205045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,1024,0.01202933324707879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,7168,0.021494223011864558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,768,0.18799111578199598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,5120,0.19998221927218965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,512,0.13961954911549887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,10240,0.7288551330566406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,65536,0.1423999998304579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,8192,0.055176887247297496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,7168,0.5460924572414821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,768,0.015105777316623263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,6144,0.02569155560599433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,8192,0.030582222673628066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,6144,0.02700533303949568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,16384,0.8718435499403211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,2560,0.04161244299676683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,2560,0.017491555876202054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,3072,0.007751999629868402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,256,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,5120,0.048335111803478666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,128,0.03921066721280416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,256,0.005352889084153705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,2560,0.008634666601816813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,256,0.009644444617960189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,8192,0.22347556220160592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,5120,0.0221742226017846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,6144,0.05359466539488899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,6144,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,6144,0.29676000277201336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,16384,0.04106222258673774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,16384,1.5781333711412218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,10240,0.02017155620786879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,512,0.020248000820477802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,1536,0.040997332996792264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,51200,0.28877332475450307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,7168,0.18733510706159803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,8192,0.012286221815480126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,51200,0.5051324632432725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,5120,0.07606933514277141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,12288,0.027275555663638647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,6144,0.06854933500289917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,1024,0.020337777005301583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,64,0.005426666802830166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,768,0.02755911151568095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,256,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,512,0.2189262178209093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,51200,1.211663140190972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,2560,0.02240444388654497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,2560,0.00850399997499254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,8192,0.9477386474609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,4096,0.07570488585366143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,2048,0.05000444584422641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,8192,0.024319110645188227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,4096,0.08215644624498156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,8192,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,16384,0.08657866716384888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,256,0.003012444410059187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,65536,1.3208292855156791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,3072,0.13542666700151232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,128,0.0028817777832349143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,8192,1.6962871551513672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,5120,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,16384,0.04237688912285698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,16384,0.02111377815405528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,6144,0.08223733637068006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,3072,0.011411555939250521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,512,0.004466666529575984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,12288,3.7019085354275174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,1024,0.009854222337404886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,2560,0.006001777946949005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,32,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,3584,0.31871644655863446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,10240,0.1698791053560045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,768,0.0277839998404185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,7168,0.01774311065673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,51200,0.42310667037963867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,3584,0.523872004614936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,768,0.004646222210592694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,5120,0.04047288828425937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,2560,0.01275644451379776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,768,0.01701866587003072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,5120,0.05561244487762451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,10240,0.026211554805437725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,51200,0.8634933365715874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,12288,0.015160888433456421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,8192,0.11909155050913493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,6144,0.026935110489527386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,128,0.009198221895429824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,256,0.006095110956165526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,8192,0.03350133366054959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,65536,0.3922871218787299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,16384,0.19034400251176622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,8192,0.016530666086408827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,5120,0.3555626604292128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,65536,2.5825778113471136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,3584,0.13378577762179905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,32,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,1024,0.0058515556156635284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,5120,0.028425776296191748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,12288,0.04484088884459602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,5120,0.014058666096793281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,3584,0.025399999486075506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,3072,0.009038222332795462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,8192,0.033017777734332614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,10240,0.22824178801642525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,3072,0.009333333207501305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,3072,0.00639111093348927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,16384,0.0365626679526435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,512,0.004815111143721474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,32,0.009827555881606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,65536,0.15660888618893093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,16384,0.20829421944088408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,3584,0.1636479960547553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,7168,0.4003777768876817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,6144,0.04915822214550442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,3584,0.09735466374291314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,768,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,8192,0.026866666144794885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,1024,0.025555556019147236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,8192,0.027221333649423387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,128,0.004920000003443824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,6144,0.2610222233666314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,51200,0.06378577815161811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,2048,0.0864408877160814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,1024,0.009126221968068017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,65536,0.08182755443784925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,768,0.005864000154866113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,64,0.030896888838873968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,16384,0.040072888135910034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,12288,0.15714755323198107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,5120,0.04081599911053976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,1536,0.00629688882165485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,256,0.005561777700980504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,4096,0.025222222010294598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,1024,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,5120,0.019721777902709115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,128,0.019314666589101154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,7168,0.01608088943693373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,5120,0.05654578076468574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,65536,0.0860675573348999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,51200,0.4000284406873915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,10240,0.01126933346192042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,10240,0.06811110840903388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,5120,0.012520888613329994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,51200,2.912913852267795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,3072,0.008709333009190029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,10240,0.07153955433103773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,1536,0.0064444442590077715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,4096,0.13385955492655435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,7168,0.01962311069170634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,51200,1.059901343451606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,16384,0.27054399914211696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,8192,0.02527644402450985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,8192,0.14789599842495388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,5120,0.08215733369191487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,2560,0.09183910820219253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,12288,0.29793866475423175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,2560,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,32,0.011384888655609555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,7168,0.23036355442470977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,6144,0.07807644208272298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,768,0.028880887561374243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,3072,0.06020977762010363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,16384,0.0635511080423991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,512,0.013568888935777875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,12288,0.12304533852471246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,8192,0.011830222275522021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,128,0.13962311214870876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,3072,0.0835804475678338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,7168,0.021191999316215515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,1024,0.006325333482689328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,8192,0.03164800008138021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,7168,0.01386844449573093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,2048,0.007272000114123027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,8192,0.3311519887712267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,12288,0.1844773292541504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,3072,0.010199999643696679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,6144,0.013001778059535556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,65536,0.2476275497012668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,4096,0.018022222651375663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,768,0.027097778187857732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,6144,0.018000889155599806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,768,0.014725334114498563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,10240,1.043701383802626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,5120,0.01127644462717904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,768,0.24015556441413033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,2560,0.1351004441579183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,51200,0.025573333104451496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,768,0.007435555259386699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,51200,14.750774807400175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,6144,0.27193066808912486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,256,0.04002666804525588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,6144,0.14878043863508436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,3072,0.009168888959619734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,256,0.0047191112405723995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,768,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,6144,0.006703111032644908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,512,0.016752888758977253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,2560,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,4096,0.042500442928738065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,768,0.004517333375083076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,1536,0.006504000061088138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,512,0.013418667018413544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,7168,0.010762666662534079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,2048,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,2048,0.12696266174316406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,2560,0.008565333154466417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,1024,0.005400889035728242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,65536,0.0813537769847446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,5120,0.009330666727489894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,4096,0.03249955508444045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,1536,0.326636446846856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,2048,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,128,0.007101332975758447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,12288,0.036278221342298723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,16384,0.02153066628509098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,2560,0.009587555295891231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,1024,0.007496000164084964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,64,0.01089511149459415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,10240,0.29855910936991376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,16384,0.11540710926055908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,51200,0.056735111607445605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,1024,0.051648888323042125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,7168,0.03383200036154853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,4096,0.06358577807744344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,6144,0.027274666561020747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,256,0.003132444495956103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,3584,0.10174755255381267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,2560,0.01908888916174571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,7168,0.9325911204020182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,1024,0.006343110981914733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,64,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,1024,0.0074480002125104266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,512,0.0037955554823080697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,3072,0.06574044624964397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,1024,0.01753155555990007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,7168,0.00887822194231881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,65536,0.2339279916551378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,768,0.00360000009338061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,64,0.006566222343179915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,65536,0.2503697872161865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,8192,6.172028435601129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,51200,0.7456622123718262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,7168,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,256,0.03897777862019009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,2048,0.00962933318482505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,65536,0.14128177695804173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,256,0.005728000154097875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,32,0.005437333136796951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,512,0.018075555562973022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,65536,0.37536001205444336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,512,0.011399110986126794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,3584,0.06344266732533772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,32,0.010481778118345471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,32,0.004670222186379963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,10240,0.01804266704453362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,32,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,4096,0.013850666582584381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,32,0.0031795555518733132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,5120,0.07450933588875665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,7168,0.27086933453877765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,6144,0.12984355290730795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,6144,0.07836088869306776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,10240,0.020008888509538438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,3584,0.021378666162490845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,1024,0.022701332966486614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,6144,3.6014552646213107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,6144,0.053370667828453906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,256,0.03958755400445726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,128,0.007436444362004598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,1024,0.004576888763242298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,768,0.01204533295498954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,2048,0.007807999849319458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,512,0.008272888759771982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,7168,0.008954666554927826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,512,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,6144,0.04968177941110399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,3072,0.01143200033240848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,3584,0.012866666747464074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,768,0.00723555518521203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,64,0.0034426665968365143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,8192,0.039825777212778725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,32,0.003087999920050303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,12288,0.34205065833197695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,5120,0.02299555473857456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,6144,0.03197688857714335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,16384,0.025042666329277888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,2560,0.021783999270863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,1024,0.010064889159467485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,32,0.0041964443193541635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,3584,0.015543111496501498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,32,0.0037288889288902283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,5120,0.3630444473690457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,7168,0.011353777514563667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,7168,0.03241689006487528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,51200,0.11394933859507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,128,0.0069013333155049225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,32,0.006663110935025745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,7168,0.016047999262809753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,512,0.025055110454559326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,51200,1.59642579820421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,7168,0.02588266630967458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,7168,0.19942577679951987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,1536,0.02108000053299798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,5120,0.023511111736297607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,1536,0.05515999926461113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,128,0.030485332012176514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,7168,0.14917155106862387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,5120,0.02444800072246128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,3584,0.0783804456392924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,2048,0.04179911149872673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,256,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,3072,0.009165333376990424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,1024,0.056030220455593534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,768,0.03206133180194431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,3584,0.008893333375453949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,16384,0.023001778456899855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,7168,0.02150933278931512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,1536,0.006558222075303395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,128,0.0040702223777771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,8192,0.041422221395704485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,10240,0.20046577188703749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,7168,0.21999200185139975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,6144,0.03587288988961114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,65536,0.021503110726674397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,1024,0.04318755533960131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,512,0.03206400076548258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,16384,0.5727555486891005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,32,0.0032879999942249725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,65536,2.308351092868381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,4096,2.1356959872775607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,64,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,16384,3.5270267062717013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,32,0.15881778134240046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,3584,0.04797955685191684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,3072,0.013462222284740873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,5120,0.08740355571111043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,7168,0.21530844105614555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,768,0.005091555416584015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,16384,0.06327199935913086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,12288,0.32543643315633136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,3072,0.017851556340853374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,2048,0.6563804414537218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,32,0.06066133578618368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,12288,0.06933422221077813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,12288,0.5290960205925835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,2048,0.07541955841912164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,6144,0.013902222116788229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,8192,0.050323555866877236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,3072,0.006002666635645761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,12288,0.6181066830952963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,65536,0.14592711130777994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,4096,0.007103111180994246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,1024,0.0806879997253418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,256,0.011483555866612328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,256,0.007905777957704332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,7168,0.1052737765842014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,10240,0.014223999447292753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,2048,0.010632889138327705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,32,0.004511111312442356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,3584,0.177727116478814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,512,0.0035866668654812705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,3584,0.007167110840479533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,12288,0.05691466728846232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,65536,2.5733155144585504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,65536,0.5254240036010742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,10240,0.10658578077952068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,3072,0.009114666945404476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,51200,0.7992489073011609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,6144,0.05562311410903931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,16384,0.006809777683681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,1024,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,512,0.007691555553012424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,256,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,5120,0.14459289444817439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,12288,0.008166222108734978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,5120,0.15316000249650744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,51200,0.11718400319417317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,2560,0.21333332856496176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,32,0.004317333300908406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,1024,0.1968639956580268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,2560,0.01405244403415256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,768,0.007189333438873291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,51200,0.16078666845957437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,16384,0.006904888898134232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,128,0.00869599978129069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,2048,0.006528888725572162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,16384,0.1833262178632948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,16384,0.02276888820860121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,1536,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,4096,0.16781866550445557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,768,0.02350577712059021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,5120,0.011642666326628791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,5120,0.13864533106486002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,1536,0.0058453331391016645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,1024,0.04612000121010674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,10240,0.3360746701558431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,768,0.006220444622966979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,3072,0.02908888790342543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,256,0.012717333104875354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,128,0.05301422211858961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,1024,0.027577777703603108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,3584,0.025760889053344727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,3072,0.35668089654710555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,768,0.030583110120561387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,8192,0.031160000297758315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,1024,0.010984000232484607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,128,0.004187555362780889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,16384,0.35469688309563535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,2048,0.014596444037225513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,64,0.015286222100257874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,1536,0.006163555714819167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,51200,0.2529519928826226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,2048,0.05301600032382541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,1024,0.13263111644321018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,3584,0.01883288886811998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,512,0.06255822049246894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,6144,0.099181334177653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,12288,0.8630764219495984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,1536,0.5651493072509766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,7168,0.03886933459175958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,65536,0.2639048894246419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,51200,0.02869510981771681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,3072,0.009377777576446533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,10240,0.05031733380423652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,16384,0.12401155630747478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,2048,0.007668444679843054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,12288,2.437355465359158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,5120,0.011365333365069495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,1024,0.00488800017370118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,5120,0.10009421904881795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,512,0.01368444412946701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,10240,0.7762728797064887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,4096,0.0218248897128635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,8192,0.0846337808503045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,2560,0.4997768931918674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,3584,0.01741155485312144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,128,0.013763555222087435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,3072,0.008902221918106079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,10240,0.02048711147573259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,16384,0.14407377772861057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,512,0.0033111110743549135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,64,0.003044444446762403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,1024,0.00434933313065105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,4096,0.026135999295446608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,3072,0.010565333068370819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,2560,0.006294222341643439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,3584,0.01756000022093455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,768,0.4817680252922906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,2048,0.011519110865063138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,6144,0.01533511115445031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,1024,0.0625200006696913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,768,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,7168,0.09405333466000027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,12288,0.2680702209472656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,8192,0.015491555134455362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,768,0.006232000059551663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,12288,0.04809066653251648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,768,0.013388444152143268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,8192,0.16264355182647705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,64,0.03991466760635376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,7168,6.070184071858724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,65536,0.3692915439605713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,7168,0.02735288937886556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,4096,0.02077866593996684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,768,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,1536,0.03725422090954251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,4096,0.1650435527165731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,51200,1.2087750964694552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,512,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,64,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,10240,0.026644445127911035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,3584,0.03479555580351088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,5120,0.018940443793932598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,1024,0.0040702223777771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,768,0.008463111188676622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,768,0.07704177829954359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,768,0.005664000080691443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,8192,1.9637376997205946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,65536,1.9201734330919054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,12288,0.8092942237854004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,512,0.030532442861133154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,768,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,51200,0.027081777652104694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,4096,0.1097244421641032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,2560,0.009215110705958473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,65536,0.14480355050828722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,256,0.006500444478458828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,65536,2.1554834577772355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,10240,0.039635555611716375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,32,0.05159822106361389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,65536,0.022965333527988855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,768,0.03046311272515191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,3072,0.007311999797821045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,5120,0.3716835445827908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,64,0.004512888689835866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,64,0.03300889001952277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,6144,0.10590310891469319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,4096,0.32502222061157227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,512,0.007713778151406183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,10240,0.17699378066592744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,2048,0.08785333236058553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,256,0.007720000214046902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,8192,0.38070933024088544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,512,0.006232000059551663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,6144,0.02743822170628442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,12288,0.02644355595111847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,1536,0.007250666618347168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,1024,0.005848000033034219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,6144,0.017287999391555786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,2048,0.02420177724626329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,16384,0.10894488626056248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,3072,0.13152266873253718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,65536,2.8998879326714406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,8192,0.008529778156015608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,128,0.00507733349998792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,64,0.008545777863926357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,8192,0.018207111292415194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,8192,0.021312000022994146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,65536,0.06544355551401775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,2048,0.012930666406949362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,1536,0.01703466640578376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,7168,0.01167022188504537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,51200,0.16823911666870117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,2560,0.023953777220514085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,256,0.023853333459960088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,8192,0.7874319818284777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,6144,0.11485600471496582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,51200,1.4890000025431316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,128,0.039247112141715154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,1024,0.007592000067234039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,3584,0.007344889144102733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,8192,0.014290667242474027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,10240,0.07358400026957194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,768,0.04176177912288242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,1536,0.01624000072479248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,4096,1.3791324827406142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,10240,0.21211022800869414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,12288,0.028432889117134943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,1536,0.04970133304595947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,8192,0.7787902090284559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,3072,0.07714044385486178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,768,0.12210933367411296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,64,0.025278222229745653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,6144,0.011220444407727985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,512,0.00350222239891688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,51200,2.107921812269423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,3072,0.03559111224280463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,768,0.006599999964237213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,1536,0.022664000590642292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,3584,0.037662221325768366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,5120,0.044827553961012095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,128,0.016201777590645682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,3584,0.0558604465590583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,512,0.0052471111218134565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,65536,1.7318844265407984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,3584,0.008886222210195329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,7168,0.013174222575293647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,256,0.016694222887357075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,3584,0.008923555413881937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,32,0.005443555613358815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,64,0.008961777720186446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,51200,0.45405954784817165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,2560,0.006981333096822103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,6144,0.028439998626708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,1024,0.012099555797047086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,3072,0.01126488877667321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,12288,0.01815200017558204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,2048,0.032590223683251276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,65536,0.0783404442999098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,1536,0.004592000196377437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,3072,0.006693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,51200,0.06260977851019965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,512,0.012858666479587555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,5120,0.025571554899215698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,51200,0.03956266575389438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,1536,0.012599111431174807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,2048,0.141412443584866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,64,0.0027866665687825945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,768,0.003535110917356279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,4096,0.415796438852946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,3584,0.17036267121632895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,64,0.006149333384301927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,1024,0.01126488877667321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,3584,0.03566577699449327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,128,0.009820444716347588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,6144,0.06279910935295953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,256,0.004230222354332606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,5120,0.0805893341700236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,6144,1.0533395343356662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,51200,3.1551867591010203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,6144,0.031163556708229914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,51200,0.16440888245900473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,1024,0.03975733452373081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,51200,0.05985866652594673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,2560,0.021599110629823472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,4096,0.16211644808451334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,6144,0.02345866627163357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,1024,0.004217777815130022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,65536,0.8351990911695691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,2048,0.006333333336644703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,512,0.025948445002237957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,768,0.033129778173234724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,4096,0.0076248885856734375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,6144,0.16001689434051514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,4096,0.011141333315107556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,8192,0.14175555441114637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,4096,0.011737777955002256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,1536,0.5017209053039551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,7168,0.01441422270403968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,64,0.004408888932731417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,5120,0.14717777570088705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,6144,3.0874862670898438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,65536,0.2252346674601237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,3072,0.016975111431545682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,64,0.008551110823949179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,256,0.0035253332720862497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,768,0.003951999876234266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,128,0.004300444490379757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,64,0.005759999983840519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,3072,0.0129439996348487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,7168,0.3576586776309543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,12288,0.18057333098517525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,2560,0.028232001596026953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,6144,0.11528089311387803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,51200,0.054429332415262856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,8192,0.023195554812749226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,512,0.852384885152181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,4096,0.01371911085314221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,4096,0.010101333260536194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,10240,0.05439111259248522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,10240,0.039769778649012245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,1024,0.22403200467427573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,3584,0.018408889571825664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,2048,0.005978666659858491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,128,0.012872888810104795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,768,0.010607110957304636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,7168,0.37824710210164386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,32,0.003232888877391815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,128,0.016930666234758165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,2048,0.01682400041156345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,65536,0.05634044276343452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,2048,0.00792622235086229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,3584,0.021985777550273474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,65536,9.549253675672743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,65536,0.07892444398668078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,32,0.006468444648716185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,6144,0.1635342174106174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,64,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,1024,0.08190755711661445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,12288,0.2454462316301134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,768,0.01255022237698237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,2048,0.005312888986534542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,8192,0.02801333202256097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,12288,0.08121511009004381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,12288,0.10016355249616836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,1024,0.02127911150455475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,10240,0.03519466519355774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,51200,0.053114665879143604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,6144,0.01587377819750044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,8192,0.01167555567291048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,32,0.005434666656785541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,128,0.020264888803164165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,512,0.15656444761488172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,768,0.0036524443162812125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,3584,0.007440889047251807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,64,0.1269893381330702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,12288,0.031110223796632554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,256,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,8192,0.1236693329281277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,768,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,3072,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,8192,14.853725857204862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,10240,0.0441599984963735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,8192,0.2948639922671848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,1024,0.0041457778877682155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,3072,9.403362698025173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,1536,0.013795555465751223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,256,0.007911110917727152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,16384,0.09898844692442153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,3072,0.010372444159454769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,2048,0.011321778098742167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,2048,0.005962666538026597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,4096,0.06198400259017944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,1024,0.005575110928879843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,16384,0.022890667120615642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,3072,0.26000089115566677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,3072,0.034308443466822304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,12288,0.7873431311713325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,2560,0.007344889144102733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,51200,0.05741688940260145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,12288,0.019468444916937087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,2560,0.006722666737106111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,3072,0.08077333370844524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,6144,3.9169832865397134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,512,0.006879999819729064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,2048,0.035558223724365234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,512,0.017679111825095285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,2560,0.03122399912940131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,5120,0.014821334017647637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,1536,0.005559999909665849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,3584,0.011560888753996955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,3584,0.009717333647939894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,5120,0.17189955711364746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,256,0.01093066649304496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,4096,0.08925155798594157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,8192,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,16384,0.1130488846037123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,2048,0.036990222003724836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,1024,0.4879475699530707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,12288,2.002133263481988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,7168,0.241376002629598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,32,0.02016444504261017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,2048,0.009927111367384592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,2560,0.044567111465666026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,3072,0.027449778384632532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,8192,0.04622933268547058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,51200,0.21539821889665392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,5120,0.25078222486707896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,3072,0.02586488922437032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,4096,0.04507022102673849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,1024,0.02940622303220961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,16384,0.05118755499521891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,16384,0.3075511190626356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,4096,0.17328177558051217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,2048,0.007895111209816402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,6144,0.02682488825586107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,12288,0.030745777818891738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,4096,0.041111111640930176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,51200,0.013791999883121915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,512,0.003504000190231535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,8192,0.5286915567186143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,4096,0.017243555850452848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,3072,0.010295111272070143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,8192,0.013295110729005603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,3072,0.0140337778462304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,2560,0.012348444097571902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,2560,0.007079111205206976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,65536,0.18351644939846465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,4096,0.1295751068327162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,3584,0.05067200130886502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,2048,0.03196622265709771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,128,0.006084444622198741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,3072,0.2377386622958713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,768,0.0047280001971456735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,4096,0.008288000192907121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,4096,0.01901333365175459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,3072,0.011586666935020022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,6144,0.04880533284611172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,128,0.0031386667655573953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,7168,0.2071768840154012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,2560,0.006239111224810283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,512,0.0086986662613021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,2048,0.008419555922349295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,3072,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,16384,0.05402222275733948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,32,0.002691555561290847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,128,0.0033235556135574975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,3584,0.018945778409639995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,3584,0.0069466663731469055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,512,0.015456888410780164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,256,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,32,0.0034906665484110513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,8192,0.042103111743927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,10240,0.022434666752815247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,4096,0.008557333714432186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,4096,0.013591999808947245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,1024,0.03866399990187751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,1024,0.01351999988158544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,512,0.0034808889031410217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,128,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,128,0.003350222276316749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,51200,0.17695377932654488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,32,0.0028808888875775864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,1536,0.00904977818330129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,6144,0.03265422251489427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,7168,0.02536711096763611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,1024,0.025314667158656653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,10240,0.028866665230857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,10240,5.999040815565322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,4096,0.010760889285140567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,10240,0.035517334938049316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,8192,0.023899555206298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,6144,0.04945688777499729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,4096,0.014114666316244336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,32,0.0033093332830402586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,64,0.005330666485759947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,32,0.006170666466156642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,3072,0.008023111356629265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,2048,0.02053333322207133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,64,0.0026586666289303037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,5120,0.14999022748735216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,16384,1.9143956502278645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,5120,0.028075555960337322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,6144,0.07064711385303073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,10240,0.1066897776391771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,6144,0.020962667134073045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,5120,0.01033866653839747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,51200,0.1446346706814236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,12288,0.01772177716096242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,128,0.008674666285514832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,8192,0.025490666429201763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,51200,2.1497376759847007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,12288,0.41062132517496747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,8192,0.009579555855857002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,64,0.014800889624489678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,7168,3.077288097805447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,512,0.06954222255282931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,768,0.14432266023423937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,32,0.0029279999434947968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,5120,0.03224710954560174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,10240,0.11189599831899007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,512,0.027689778142505225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,10240,0.08313955863316853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,5120,0.0170168893204795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,256,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,3072,0.010297777752081553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,5120,0.05283022258016798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,16384,0.051001777251561485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,1536,0.0819022191895379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,6144,0.3005466726091173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,32,0.0076657773719893555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,2048,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,6144,0.01314755529165268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,3584,0.009226666556464301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,768,0.03424177898301019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,768,0.035880890157487653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,7168,0.013036444783210754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,2560,0.02088622252146403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,768,1.0164835188123915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,12288,2.9970648023817272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,64,0.006509333435032103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,8192,0.3349413341946072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,2560,0.008839999636014303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,8192,0.04026666614744399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,2048,0.009677333136399588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,32,0.0034888887570963967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,128,0.005903111149867375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,64,0.0033377777371141645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,5120,0.023445333043734234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,512,0.02327288852797614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,512,0.006706666615274217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,4096,0.18711288770039877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,768,0.013359111216333179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,3584,0.01718133356836107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,65536,0.5082329114278158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,65536,0.2927030987209744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,1024,0.0063795554969045855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,65536,2.5955448150634766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,64,0.013122667041089801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,768,0.007183999651008182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,8192,0.3340764575534397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,64,0.06422399812274508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,1536,0.045850667688581676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,10240,0.21053955290052626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,51200,0.061200890276167125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,16384,0.07756711377037896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,12288,9.919358147515192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,128,0.009526222116417354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,5120,0.3042248884836833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,8192,0.025435555312368605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,7168,3.1805422041151257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,2048,0.01199288914601008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,12288,8.09042273627387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,5120,0.03624533282385932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,1024,0.1633173359764947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,64,0.05316355493333605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,32,0.005099555684460534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,4096,0.031152000029881794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,512,0.020310221446885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,3584,0.006703111032644908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,3072,0.0778097775247362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,1536,0.0069191112286514705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,3584,0.1266382270389133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,8192,0.011888888974984487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,8192,0.14019644260406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,2560,0.020421332783169217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,65536,0.1504462162653605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,2560,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,1536,0.011776888536082374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,10240,0.1628053320778741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,4096,0.05829333596759372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,3072,0.1825502183702257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,12288,0.11213333076900905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,10240,0.32821867201063365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,6144,0.05369155605634054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,1536,0.00960355583164427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,3584,0.1254835526148478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,16384,0.01827555563714769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,64,0.017457778255144756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,8192,0.012172444827026792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,2048,0.0058853332367208265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,128,0.010192888478438059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,1024,0.009372444616423713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,3584,0.4179191059536404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,1536,0.007786666353543599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,5120,0.02160088883505927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,1024,0.015676443775494892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,512,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,10240,0.03293155630429586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,51200,0.5751528739929199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,6144,0.012818666795889536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,1536,0.013816888961527081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,51200,20.65500725640191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,768,0.019084443648656208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,512,0.009918221996890174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,256,0.0031591111587153543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,64,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,128,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,7168,1.3387991587320964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,768,0.012551111479600271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,1024,0.0653102199236552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,768,0.006976000136799282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,12288,0.019176888797018263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,16384,0.041377776198916964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,512,0.004882666799757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,7168,0.4031679895189073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,2560,0.016939555605252583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,1024,0.029639111624823675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,65536,0.3597911198933919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,6144,0.03128977616628011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,2560,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,4096,0.019156444403860304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,1024,0.006730666591061487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,5120,0.037674665451049805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,3584,0.008750222623348236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,65536,0.06422577963935004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,10240,0.5764808654785156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,3584,0.006430222342411677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,10240,0.23804977205064562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,3072,1.2849946551852756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,12288,0.05571644173728096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,65536,0.07685599724451701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,6144,0.5414826605055068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,2048,0.07396711243523492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,32,0.008616888688670265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,128,0.003395555540919304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,768,0.008720888859695857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,512,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,7168,0.23756355709499785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,7168,0.025216889050271776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,768,0.031089779403474595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,51200,0.31135556432935924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,8192,0.061046222845713295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,4096,0.011334222224023608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,1536,0.2926017708248562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,2048,0.005802666561471091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,7168,0.01719911065366533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,3584,0.22423111067877877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,6144,0.1317173375023736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,4096,5.020113627115886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,5120,0.20709778202904594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,6144,0.01922044489118788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,4096,0.04901866780387031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,3072,0.09993422031402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,2560,0.014198222094111972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,8192,0.09091999795701768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,7168,0.30398933092753094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,1536,0.01735733283890618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,7168,0.10619199938244289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,1024,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,16384,0.5483324262830946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,1024,0.007091555330488417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,2048,0.006791111081838608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,5120,0.013376889129479727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,128,0.0033164444482988785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,12288,0.061905774805280894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,2560,0.019271110494931538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,768,0.004699555536111196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,128,0.004536889079544279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,12288,0.41346221499972874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,1024,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,4096,0.276293330722385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,3072,0.11660444736480713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,3584,0.00795733349190818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,1024,0.006392000036107169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,2560,0.022318222456508215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,1536,0.009919111099508073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,7168,0.017323555217848886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,768,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,16384,0.47814310921563047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,12288,0.03502310978041755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,4096,0.009275555610656738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,51200,0.33972711033291286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,7168,0.011003555523024665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,3072,0.0738453335232205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,32,0.00997688869635264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,128,0.18291289276546904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,128,0.014521777629852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,5120,0.06350488795174493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,768,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,64,0.012572444147533841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,51200,15.843044704861112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,6144,0.019506666395399306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,5120,0.1633333365122477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,6144,0.011360888679822287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,64,0.005727111051479976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,16384,0.04336533281538221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,1536,0.061392002635531955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,10240,2.2521190643310547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,2560,0.04286844531695048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,4096,0.01388177772363027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,2560,0.07394666804207696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,12288,0.016128889388508268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,1536,0.01662488944000668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,512,0.00573244442542394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,3584,0.025993777645958796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,2048,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,1536,0.8282257715861002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,12288,0.1760622262954712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,1024,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,12288,0.01938222183121575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,65536,0.07474044296476576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,51200,0.06851644648445977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,16384,0.2572817802429199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,64,0.010220444036854638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,8192,0.020459555917316012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,5120,0.09114666779836018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,2560,0.007684444387753804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,512,0.025144888295067683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,3584,0.010865777730941772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,32,0.003043555551105075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,10240,0.265521764755249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,3584,0.39716800053914386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,512,0.003580444388919406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,128,0.014737778239780001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,16384,0.11359643936157227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,3072,0.009482666850090027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,256,1.2697920269436307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,10240,0.4887670940823025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,51200,0.01330311099688212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,8192,0.04488889045185513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,768,0.02292177743381924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,512,0.020576889316240948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,128,0.035897778140174016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,64,0.0063928887248039246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,64,0.005817777580685086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,768,0.00904622260067198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,3072,0.0843893355793423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,64,0.005223999834722943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,128,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,3584,0.015996444556448195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,2048,0.010952889091438718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,64,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,768,0.0035617777870761026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,6144,0.27622487809922963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,1536,0.01635199950801002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,3072,0.010955555571450127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,3584,0.14710488584306505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,8192,0.0119991112086508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,12288,0.020111999577946134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,5120,0.03083555565940009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,2048,0.02124888863828447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,3584,0.026455110973782007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,128,0.004930666751331753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,2048,0.017249777913093567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,51200,1.6455137464735243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,32,0.0027306667632526825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,32,0.003170666595300039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,512,0.00452533322903845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,3584,0.008332444561852349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,768,0.01920711166328854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,64,0.005316444569163852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,2048,0.02460533380508423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,7168,0.15855822298261854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,768,0.014725334114498563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,10240,0.007081777685218387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,51200,0.13203822241889104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,2048,0.02624711063173082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,1024,0.012907555533779992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,12288,0.04500977860556709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,2048,0.2991093264685737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,32,0.0030577776746617425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,10240,0.030496888690524634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,768,0.006127999888526068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,1536,0.0058577776783042485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,6144,0.01443733274936676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,1536,0.10806755224863689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,256,0.09541155232323541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,256,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,64,0.0028533333291610083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,4096,0.011633777783976661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,128,0.008135110967689091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,10240,0.34277600712246364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,10240,0.16832533147599962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,5120,0.009940444595283931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,5120,0.02032355633046892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,2048,0.02683199942111969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,4096,0.22062844700283477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,3072,0.015328889091809591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,128,0.013362666798962487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,3072,0.060864885648091636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,3072,0.018783110711309645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,6144,0.10182399882210626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,128,0.004860444615284602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,32,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,64,0.004022222426202563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,512,0.005558222118351195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,64,0.006960888703664144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,8192,0.18151377307044136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,65536,0.25879022810194224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,3072,0.025387555360794067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,1024,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,65536,1.6553608576456706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,64,0.0046364445653226645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,128,0.0031831111345026228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,1024,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,1024,0.051551110214657254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,65536,0.21031644609239367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,65536,0.07248977820078532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,256,0.010775999890433418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,3072,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,10240,0.34194310506184894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,16384,0.07389511002434625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,4096,0.022437334060668945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,3072,0.01735555628935496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,512,0.008820444345474243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,32,0.11759732829199897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,3584,0.011332444846630096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,6144,0.011884444289737277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,51200,0.5519182417127821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,7168,0.5021831194559733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,10240,0.1503395504421658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,2048,0.010085333552625444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,64,0.014871999621391296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,10240,0.042949332131279834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,128,0.0027911110470692315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,1024,0.006186666587988536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,4096,0.014263111684057446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,3584,0.2506328953637017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,1536,0.06020977762010363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,51200,0.19302311208513048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,768,0.0058168888919883305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,3584,0.6749351289537219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,16384,0.02201777696609497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,256,0.012299555871221753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,128,0.0034737777378824023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,256,0.0052284445199701525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,5120,0.015358222855461968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,51200,0.02608088817861345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,2560,0.014352000421947904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,32,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,10240,0.046353777249654136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,6144,0.010068444742096795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,64,0.002921777880854077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,7168,0.02819288770357768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,64,0.011008888483047485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,4096,0.012142221960756512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,1536,0.014440000057220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,3584,0.04417777723736233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,65536,0.0690222250090705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,2560,0.009563555320103964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,128,0.008093333078755273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,32,0.005161777966552311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,3584,0.04093155595991347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,3072,0.01255644443962309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,2560,0.04683822062280443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,2048,0.03489866521623399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,256,0.003085333233078321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,65536,0.0943982203801473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,8192,0.09884622361924912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,768,0.0038382220599386427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,1024,0.00756533361143536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,5120,0.01897333396805657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,10240,0.01889244384235806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,512,0.005515555540720622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,16384,0.10522933138741387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,12288,0.6300151083204482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,2560,0.03793333305252923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,4096,0.010636444720957013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,5120,0.016170667277442086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,2048,0.008480889101823172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,1024,0.007855999800893996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,64,0.0037395556767781577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,7168,0.06384178002675374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,2560,0.008928000099129146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,1536,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,10240,0.26532088385687935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,2048,0.0116462219092581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,32,0.17699200577206084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,768,0.0321048895517985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,768,0.004616888860861461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,2048,0.00693866651919153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,1024,1.5252293480767145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,6144,0.08480088578330146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,3072,0.08068711227840848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,2048,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,4096,0.11085777812533909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,4096,0.008106666306654612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,6144,0.020841777324676514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,512,0.005534222142563925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,2048,0.006215999937719769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,12288,0.3400293456183539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,2048,0.020920000142521326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,12288,0.1927271154191759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,4096,0.016582222448454965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,51200,4.439330630832248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,65536,0.3042702145046658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,2560,0.019496000475353666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,512,0.0162453336848153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,4096,0.007207999626795451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,128,0.011743110915025076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,6144,0.8500284618801541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,64,0.14582310782538518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,5120,0.11063022083706325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,6144,0.035424888134002686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,64,0.002990222225586573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,7168,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,10240,0.07970399989022149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,10240,0.03222577770551046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,5120,0.14583644602033827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,5120,0.009088888764381409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,512,0.0038035557501845886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,8192,0.2683066791958279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,1536,0.03359288970629374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,5120,0.011651555697123209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,7168,0.03602577911482917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,768,0.032878223392698504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,3584,0.06669155756632487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,2048,0.023539554741647508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,6144,0.01832799944612715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,65536,0.05630489190419515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,3072,0.010658666491508484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,51200,0.06283199787139893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,3584,0.009319110876984065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,1024,0.005897777775923411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,6144,0.1975733306672838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,1024,0.005768888940413793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,768,0.03641066617435879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,7168,0.022376888328128394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,65536,0.18741956022050646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,2560,0.007124444676770105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,3072,0.01565511193540361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,3584,0.31557777192857533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,3072,0.14839111434088811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,32,0.0036675557494163513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,3072,0.01998222205373976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,3072,0.017151110702090792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,2560,0.008262222011884054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,5120,0.22809243202209473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,5120,0.06303466690911187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,128,0.005487111293607288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,32,0.0026586666289303037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,256,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,256,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,2048,0.06975911060969035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,8192,0.01182933317290412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,512,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,2048,0.4288586775461833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,768,0.012983111043771109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,3584,0.05023022161589729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,6144,0.6588444179958767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,256,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,5120,0.007989333735571967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,3072,1.1690773434109158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,8192,0.07647733555899726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,10240,0.019140443868107267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,128,0.004442666553788715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,51200,0.8397679858737521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,128,0.05756622552871704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,1024,0.005616888817813661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,512,0.008355555435021719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,768,0.0036559998989105225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,64,0.024103111690945093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,6144,0.1628968848122491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,3072,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,256,0.005517333332035277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,51200,0.17709155877431235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,7168,0.027767111857732136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,1024,0.027657777070999146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,64,0.060277336173587374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,51200,0.08271288871765137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,128,0.03177600105603536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,3584,0.009032888544930352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,12288,0.6653457747565376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,2048,0.009377777576446533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,16384,0.11478843953874375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,5120,0.02093066606256697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,3072,0.021335111724005804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,512,0.09783289167616104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,6144,0.010863111250930362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,7168,0.01690933273898231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,51200,0.4171120060814752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,3072,0.03543555405404832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,1024,0.005505777895450592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,10240,0.17407643795013428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,128,0.011305777562989129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,3584,0.03143999973932902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,32,0.016846223009957206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,65536,0.08098577790790133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,512,0.0033271111961868075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,5120,0.01091644416252772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,7168,0.01369066701995002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,4096,0.03180444571706984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,2560,0.013280889226330651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,12288,0.07869778076807658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,51200,0.41821689075893825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,6144,0.01370666672786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,128,0.004467555632193883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,3072,0.11226133505503337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,256,0.003793777690993415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,8192,0.026294221480687458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,2048,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,32,0.002672000063790215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,2048,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,16384,0.023652444283167522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,128,0.006979555719428592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,12288,0.040186666780047946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,10240,1.4265421761406794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,2048,0.0716080003314548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,256,0.02103288968404134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,768,0.012087110843923358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,3072,0.015086222026083203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,10240,0.7128284242418078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,2560,0.07737421989440918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,10240,0.23363023334079316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,6144,0.022089777721299067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,1024,0.012630222572220696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,2560,0.006575111299753189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,64,0.0031582222630580268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,1024,0.026572444372706946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,1536,0.017459554804695975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,8192,0.2011422183778551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,4096,0.03738311264250014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,8192,0.08732622199588352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,5120,0.009189333352777693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,1024,0.060903999540540904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,51200,0.05299999978807238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,2560,0.04078311059210036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,6144,0.020814221766259935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,12288,0.043544888496398926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,16384,1.2485137515597873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,4096,0.24714133474561903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,64,0.15939377413855657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,7168,0.057355556223127574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,32,0.006519111080302133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,65536,0.07557777563730876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,4096,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,65536,0.022252443763944838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,256,0.005593777944644292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,768,0.37605243259006077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,16384,1.329451560974121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,2560,0.0069022224181228215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,3072,0.07748177978727552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,2560,0.8439493179321289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,10240,0.08109244373109606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,16384,0.10857955614725749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,2560,0.19449955887264678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,2560,0.06494666470421685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,2048,0.02459822263982561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,32,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,32,0.0027422221998373666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,10240,0.3411404556698269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,16384,6.690234714084202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,768,0.00998933364947637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,51200,0.5151475800408257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,16384,0.542410691579183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,256,0.011167999770906238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,6144,0.28195378515455455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,65536,3.181066724989149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,5120,0.052037331793043345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,51200,0.023759111762046814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,256,0.008642666869693333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,3072,0.04627822173966301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,2048,0.007496889266702864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,4096,0.10748533407847087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,2560,0.044736001226637095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,12288,0.01703377730316586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,65536,5.8287260267469625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,3584,0.07776533232794867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,1536,0.012146666646003723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,12288,0.17909867233700225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,1536,0.012879999975363413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,32,0.003277333246337043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,7168,0.050670223103629224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,3072,0.01236355553070704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,64,0.0040702223777771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,1024,0.018212444252438016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,10240,0.010121777653694153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,6144,0.057787557442982994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,10240,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,512,0.0031386667655573953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,7168,0.04437244600719876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,51200,0.06843466891182794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,12288,0.015014222926563688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,512,0.006139555739031897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,10240,0.06486666864818998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,7168,5.818994734022353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,3584,0.033196445968416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,1024,0.004173333446184794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,5120,0.19250044557783338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,2560,0.16873866981930205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,256,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,8192,0.5308631261189779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,5120,0.5595813327365452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,16384,0.0380968881977929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,128,0.004292444636424382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,3584,0.008028444316652086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,12288,0.3061537742614746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,3072,0.10888533459769355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,16384,0.20304977893829346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,51200,0.06124622291988797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,3584,0.01868977811601427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,1024,0.026663111315833196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,1536,0.018664888209766813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,1536,0.022846221923828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,3072,0.007458666960398356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,6144,0.06313421991136339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,2048,0.05008888906902737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,10240,0.503910223642985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,10240,0.023734221855799358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,16384,0.14741688304477268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,4096,0.057633777459462486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,10240,0.02698933415942722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,1024,0.024863999750879075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,3072,0.009694221946928237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,768,0.006035555568006303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,7168,0.0513573322031233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,3584,0.6078249083624946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,65536,0.11906222502390544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,2048,0.021578666236665513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,8192,0.013334222137928009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,65536,171.6329888237847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,10240,0.021951999929216173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,3584,0.012106666962305704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,512,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,16384,0.5716622140672472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,7168,0.1513706710603502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,16384,0.44945510228474933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,768,0.006247111078765657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,16384,0.07712178097830878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,2560,0.010609777437316047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,3072,0.021376889612939622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,16384,0.019343111250135634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,4096,0.011367110742463006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,6144,0.12106400065951878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,3584,0.019905777441130746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,16384,0.1815199984444512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,1536,0.004596444467703502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,7168,0.02960977620548672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,4096,0.03779733180999756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,1536,0.011763555308183035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,4096,0.014594667487674289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,1024,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,3072,1.241487079196506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,65536,0.18666488594479033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,512,0.23650667402479383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,7168,0.025861332813898723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,1024,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,64,0.005105777747101254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,5120,4.259956359863281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,64,0.007872000336647034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,6144,0.08953244156307644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,3072,0.13519822226630315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,4096,0.15242221620347765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,5120,0.015135111080275642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,3584,0.41225422753228086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,2560,0.00959644466638565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,5120,0.015623999966515435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,2048,0.017295110556814406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,8192,0.014908444550302295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,51200,1.2308746973673503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,2560,0.7887360254923502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,2560,0.016386666231685214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,64,0.0034764446318149567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,7168,0.3501617908477783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,3072,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,1024,0.00620888877246115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,256,0.007063999772071838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,1536,0.009575111170609793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,2560,0.005945777727497949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,128,0.04820000131924947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,12288,0.07438221904966566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,256,0.01961511042382982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,4096,0.1894293361239963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,768,0.009892444643709395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,51200,0.13508889410230848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,10240,0.029662221670150757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,32,0.01603200038274129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,7168,0.05586489041646322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,1024,0.005468444277842839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,256,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,2560,0.06679733594258626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,6144,0.03670488794644674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,3584,0.10981066359413995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,51200,0.553999106089274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,2560,0.028811554114023846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,1024,0.006884444504976273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,7168,0.09953688912921482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,16384,0.17589867115020752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,12288,0.03611466619703505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,3584,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,256,0.012081777883900536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,4096,0.01770400007565816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,32,0.0029644444584846497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,3072,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,256,0.013186666700575085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,5120,0.01724088854259915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,10240,0.3193804423014323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,16384,0.8685199949476453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,64,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,65536,0.4447982046339247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,16384,0.006063999815119638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,1024,0.0069777775141927935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,1024,0.021026665965716045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,65536,1.3789342244466145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,512,0.014726221561431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,1536,0.05628799729877048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,768,0.03226666649182638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,7168,0.014069333672523499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,51200,0.6307279798719618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,2560,0.03469244307941861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,12288,0.04026666614744399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,1536,0.01679022279050615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,512,0.006121777825885349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,768,0.006196444647179709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,1024,0.00831200016869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,65536,0.25335555606418186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,2560,0.006241777704821692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,3584,0.02085777786042955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,1536,0.0079644446571668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,4096,0.10736711157692803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,128,0.018536888890796237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,512,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,512,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,2048,0.07614399989446004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,51200,0.04961599906285604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,3072,0.10097955332862006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,32,0.006559999866618051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,65536,0.29461778534783256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,768,0.007821333077218797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,2048,0.02993511160214742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,51200,0.008512000242869059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,16384,0.07787377966774835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,4096,0.10991289218266805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,7168,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,3072,0.015190222197108798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,4096,0.03121688961982727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,3072,0.07659733295440674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,128,0.018965333700180054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,8192,0.013839110732078552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,512,0.035247110658221774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,3072,0.006968888971540663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,2048,0.23379998736911348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,65536,0.1456648906071981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,6144,0.028232001596026953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,1536,0.024837333295080397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,6144,0.006795555353164673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,512,0.007328888608349695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,16384,0.047003553973303906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,768,0.003875555677546395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,1536,0.006976889239417181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,7168,0.01422755585776435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,1536,0.009135111338562435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,2560,0.026936888694763184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,4096,0.42184800571865505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,3584,0.008346666892369589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,32,0.005767999837795894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,3072,0.007594666547245449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,256,0.2654160128699409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,2048,0.11552266279856364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,65536,0.11174666881561279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,6144,0.014959999256663851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,3072,0.0199955552816391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,32,0.03368977705637614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,2560,0.7395360204908582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,65536,0.1731875605053372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,128,0.0029235554652081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,1024,0.07671022415161133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,4096,0.020228443874253165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,8192,0.3460933367411296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,12288,0.0247697780529658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,4096,0.14694044325086805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,2560,0.013311111264758639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,8192,0.014183110660976835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,12288,0.0321777794096205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,8192,0.19634666707780626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,3072,0.05258488986227247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,768,0.0047191112405723995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,2560,0.012844444149070315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,10240,0.06033600038952298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,64,0.0035235554807715942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,65536,0.19767733414967856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,6144,0.1831493377685547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,2560,0.00815288888083564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,12288,0.08297689093483819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,5120,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,3072,0.006490666833188798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,65536,3.632733239067925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,5120,0.07009689013163249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,512,0.031587554348839655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,2560,0.02678044471475813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,4096,0.020827554994159274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,65536,1.2360622617933485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,8192,0.12002400557200114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,64,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,51200,2.1734434763590493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,256,0.003647110942337248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,16384,0.46118757459852433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,16384,1.2214987013075087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,2560,0.015388444066047668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,128,0.003798222376240624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,51200,1.3454898198445637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,4096,0.00997688869635264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,6144,0.2187324497434828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,3072,0.009611555271678502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,1024,0.007962666451931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,2560,0.009767110976907942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,7168,0.015445333388116626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,16384,0.044723557101355664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,8192,0.38377777735392254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,512,0.006215999937719769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,768,0.007087111473083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,3584,0.06212177541520861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,10240,4.745062086317274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,6144,0.018191999859280057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,12288,0.0958942241138882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,512,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,12288,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,1024,0.07115555471844144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,32,0.0036186666952239144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,51200,0.043112887276543506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,65536,0.010950222611427307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,768,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,7168,0.17569955190022787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,3072,0.007988444632954067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,7168,0.015582222077581616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,3072,0.06146488587061564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,12288,0.018983110785484314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,5120,0.012860444684823355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,51200,0.15369778209262425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,3072,0.954793824089898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,8192,0.38293690151638454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,3072,0.015944888194402058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,3072,0.007827555967701806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,768,0.006385777973466449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,6144,0.1362640062967936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,3584,0.009509333305888707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,2048,0.013942221800486246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,3584,0.04397777716318766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,256,0.004361777669853634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,4096,0.026720000637902155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,5120,0.013400889105266996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,12288,0.125763561990526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,64,0.00980977796845966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,51200,0.055995557043287486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,64,0.1592302189932929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,512,0.03166666626930237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,51200,0.09948088725407918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,2560,0.007222221957312689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,32,0.027871999475691054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,3072,0.9927298227945963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,1024,0.019303111566437613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,768,0.026192888617515564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,32,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,256,0.0069431112044387394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,512,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,65536,31.068745930989582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,768,0.006867555694447623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,256,0.01293333371480306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,12288,0.02940888868437873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,8192,5.500495062934028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,8192,0.07421955797407362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,3584,0.13579466607835558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,3072,0.018159111340840656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,2560,0.08856888612111409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,256,0.005435555759403441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,8192,0.034210665358437434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,2560,0.00960000024901496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,256,0.01827822294500139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,16384,0.4209706783294678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,256,0.003306666596068276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,10240,10.316329108344183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,6144,0.1870737738079495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,128,0.03936444388495551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,256,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,3584,0.03979111048910353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,128,0.003048000029391713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,65536,0.29531910684373647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,12288,0.029179556502236262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,512,0.008850666383902231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,6144,0.009927111367384592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,128,0.0047733332547876565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,1024,0.014135110709402295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,7168,0.021428444319301184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,65536,0.07029244634840223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,768,0.08075111442142062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,51200,0.051631109582053296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,1024,0.1196240054236518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,512,0.005994666781690385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,768,0.04229155513975355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,16384,0.0269768883784612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,3072,0.026480888326962788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,16384,0.021224000387721594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,3072,0.37311556604173446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,2560,0.28236177232530385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,7168,0.09520710839165582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,65536,0.11607110500335693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,10240,1.5309279759724934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,32,0.00327822214199437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,65536,5.494430118136936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,7168,0.29048177931043834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,512,0.005352889084153705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,6144,0.04366933306058248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,5120,0.0074337778819931885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,512,0.057483553886413574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,51200,25.808817545572918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,6144,0.010584000084135266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,768,0.43107112248738605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,1536,0.006704000135262807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,7168,0.031692445278167725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,3072,0.038704001241260104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,256,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,10240,0.04194400045606825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,64,0.0028524444335036804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,1536,0.05254400107595655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,32,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,3072,0.008651555412345463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,512,0.010570666856235929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,1536,0.00923288861910502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,10240,0.026124444272783067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,64,0.002720000015364753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,128,0.007995555798212687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,64,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,3072,0.013785777820481194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,128,0.007164444360468123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,2560,0.7513582441541883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,8192,0.03643644518322415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,8192,0.16349955399831137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,32,0.004857777721352047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,64,0.009287111461162567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,1536,0.07810311184989081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,128,0.008994667066468133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,16384,0.024911110599835713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,1024,0.006812444577614467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,51200,3.473579618665907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,4096,0.041851556963390775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,2048,0.007205333146784041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,3072,0.011221333510345884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,6144,0.02416533397303687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,2048,0.591208881802029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,256,0.08551022079255845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,3584,0.008101333346631791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,3072,0.07524710893630981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,3584,0.01739911072784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,5120,0.020248888267411124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,256,0.034288889831966825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,5120,0.016089777151743572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,32,0.06199555264578926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,3584,0.03030666708946228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,4096,0.010246222217877706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,256,0.018343999981880188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,32,0.004582222137186262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,16384,0.1719342205259535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,10240,0.5572097566392686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,12288,0.2664097679985894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,768,0.00906311141120063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,4096,0.01663733356528812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,64,0.012553777959611682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,2560,0.01366666704416275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,6144,0.06588888830608793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,7168,0.05607822206285265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,32,0.009395555489593083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,8192,0.023380445109473333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,8192,0.07165066401163737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,768,0.12180444929334854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,4096,0.012253333297040729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,65536,0.09905422396130031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,32,0.019354666272799175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,768,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,512,0.0034462221794658234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,12288,0.0414720012081994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,1024,0.004054222255945206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,768,0.00999466660949919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,3584,0.14109867148929173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,128,0.021225778592957392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,3584,0.013062222136391534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,3584,0.006989333364698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,10240,0.3042648898230659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,768,0.008827555510732863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,768,0.004039999925427967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,51200,0.1362231175104777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,2560,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,768,0.005937777873542573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,256,0.013320888910028668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,4096,0.2622240119510227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,4096,0.027689778142505225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,128,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,2560,0.07500266366534762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,10240,0.07755021916495429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,10240,0.3518826696607802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,65536,0.06632622083028157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,7168,0.01146577795346578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,768,0.0057964444988303715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,64,0.02083555526203579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,512,0.3186737696329753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,3584,0.02349511120054457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,2048,0.25023288196987575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,3584,0.052071998516718544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,7168,1.4438915252685547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,65536,0.05866222249137031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,1024,0.0053582220441765254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,4096,0.06415733363893297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,512,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,32,0.0026542221506436667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,5120,0.04218222366438972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,2560,0.11050666703118218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,65536,0.1457564433415731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,512,0.09009422196282281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,5120,0.023677332533730402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,12288,0.059557331932915576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,3072,0.03608977794647217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,12288,0.11055999994277954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,8192,0.018562666243977018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,768,0.007720000214046902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,256,0.003663111064169142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,6144,0.01149777736928728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,2048,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,3072,0.08849066495895386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,3584,0.025431111454963684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,3584,0.010814222196737925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,10240,0.11049333545896743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,512,0.02500266664557987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,16384,0.11278755135006374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,3584,0.00814488861295912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,512,0.010405333505736457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,256,0.020409777760505676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,64,0.005737777799367905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,6144,0.021372444099850122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,7168,13.896183437771269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,1024,0.05247911148601108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,3072,0.0129120002190272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,768,0.016051555673281353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,512,0.01147377739350001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,5120,0.4347253375583225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,3072,0.07147910859849718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,16384,0.020850666695170935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,512,0.0058151111006736755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,5120,0.017457778255144756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,7168,0.0085235552655326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,10240,0.04047288828425937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,7168,0.0415013333161672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,7168,0.027758222487237718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,256,0.005251555393139522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,2048,0.03523822294341193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,256,0.0028444443725877335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,128,0.00674044465025266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,16384,0.3782648775312636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,256,0.005107555538415909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,16384,0.07715733183754815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,10240,0.4043724536895752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,4096,0.5810622109307183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,7168,0.20394578244951037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,2048,0.3008951081169976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,4096,0.013286222186353473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,8192,0.34179645114474827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,1536,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,7168,0.028139554791980322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,128,0.0028159999185138275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,768,0.016232000456915963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,2560,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,1024,0.006908444480763541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,128,0.002761777697337998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,768,0.006180444525347815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,128,0.00871733327706655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,5120,0.013360000318951078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,8192,0.01907822158601549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,2560,0.07040088706546359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,256,0.002874666617976295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,7168,1.1389492882622612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,65536,0.5680070983039008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,3072,0.008401778009202745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,2048,0.023751111494170293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,7168,0.13716799683041042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,3072,0.010392000277837118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,1024,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,12288,0.23174754778544107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,256,0.2437137762705485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,5120,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,1536,1.1049928665161133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,4096,0.03624711102909512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,7168,0.04863644308514065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,65536,0.2009119987487793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,3072,0.011971555650234222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,10240,0.1569004456202189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,4096,0.08167911238140531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,2560,0.7212417920430502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,3072,0.012543111211723752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,4096,0.013016888664828407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,128,0.007132444116804335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,3072,0.17419732941521537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,768,0.023768888579474554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,12288,0.03253777821858724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,10240,0.15392177634769016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,65536,0.06404533651140001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,10240,0.2074035538567437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,5120,0.012272888587580787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,1536,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,12288,0.07173066669040255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,16384,0.03059911065631443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,32,0.002609777781698439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,32,0.029846221208572388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,7168,0.31685511271158856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,2560,0.00627377794848548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,32,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,128,0.003917333152559069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,3072,0.011821333732869891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,64,0.0035217776894569397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,6144,0.11897244718339707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,10240,0.014768888552983602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,512,0.005616888817813661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,1536,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,1536,0.03654044535424974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,12288,0.021988444858127173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,512,0.0037466668420367767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,2048,0.0325582226117452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,2048,0.008753777378135258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,7168,0.05574488639831543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,128,0.006219555520349079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,12288,1.1718577278984916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,5120,0.045033779409196645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,6144,0.06974489159054227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,768,0.005605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,8192,0.027836445305082533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,512,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,8192,0.03038133184115092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,3584,3.0064222547743054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,256,0.007325333025720384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,2048,0.007188444336255391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,8192,0.23832355605231392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,51200,1.3163955476548936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,10240,0.25793955061170787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,32,0.009660444325870937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,2048,0.03552622265285916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,10240,0.007103111180994246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,8192,0.037017779217825994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,3584,0.19438044230143228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,51200,3.316970613267687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,768,0.018579555882347953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,7168,0.03724799884690179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,6144,2.9136435190836587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,768,0.007263110743628607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,10240,0.266482671101888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,10240,0.07066933314005534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,4096,0.1093368861410353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,12288,0.04483022292455038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,16384,0.03728977839152018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,7168,0.03923733366860284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,65536,3.30492803785536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,3584,0.0732791092660692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,512,0.00647555540005366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,3072,0.01553244392077128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,512,0.030537777476840552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,768,0.005752888818581899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,1024,0.015088000231319003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,2048,0.035214222139782377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,512,0.006416888700591193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,16384,0.009176000124878353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,3584,0.880425771077474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,51200,0.16215644942389593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,3584,0.00742222203148736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,2560,0.040920890039867826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,12288,0.14500444465213352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,3584,0.10042399830288357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,3584,0.020780444145202637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,4096,0.0627431141005622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,512,0.004271111140648524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,512,0.008845333423879411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,51200,0.16273955504099527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,5120,0.10241511132982041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,128,0.0092622223827574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,256,0.16050044695536295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,7168,0.01444355481200748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,256,0.006599111275540457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,7168,0.018750222192870248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,512,0.005462222215202119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,5120,0.018481777773963082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,3072,0.01035822182893753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,3072,0.011952000359694162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,6144,0.06944533189137776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,7168,25.139933268229168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,1024,0.010246222217877706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,51200,4.2212880452473955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,3072,0.033612443341149226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,3072,0.02911111050181919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,3072,0.009119999905427298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,2560,0.006805333412355847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,6144,0.03871466716130575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,768,0.20811466375986734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,65536,0.040103111002180315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,32,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,3072,0.015633778439627755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,3072,0.9952738020155164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,10240,0.0329022208849589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,32,0.0036142224238978494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,4096,0.018613333503405254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,128,0.041737778319252856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,7168,0.18155822488996717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,1536,0.007956444389290279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,7168,0.006375999914275275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,7168,0.018981332580248516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,12288,0.019374221563339233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,5120,0.02784622377819485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,8192,0.14894133143954807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,4096,0.016759110821617972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,65536,0.9051262537638346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,51200,0.4308906661139594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,1024,0.003940444439649582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,5120,0.10247911347283258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,2560,0.010215111076831818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,2560,0.13313955730862087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,16384,0.426804436577691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,3584,0.09462310870488484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,12288,0.147800882657369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,8192,0.06406578090455797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,5120,0.03351644343800015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,64,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,256,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,768,0.02055022286044227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,512,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,3072,0.010377777947319878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,768,0.005758222192525864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,16384,0.4980826907687717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,128,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,1536,0.010830221904648675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,10240,0.689426687028673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,12288,0.04039555456903245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,6144,0.07920000288221571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,6144,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,10240,0.02312266661061181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,8192,0.2943449020385742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,3584,0.18388711081610787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,51200,0.3280248906877306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,1536,0.3920444382561578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,51200,0.5059057871500651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,64,0.007319110963079665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,2560,0.021012443635198805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,7168,0.1301199992497762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,1024,0.006882666713661618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,12288,0.008765333228641087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,10240,0.09664444128672282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,65536,30.304331461588543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,16384,0.11141510804494222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,8192,0.031054221921496924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,256,0.11142488320668538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,5120,0.10196089082294041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,3584,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,16384,0.10299911101659139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,4096,0.018388445178667705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,5120,0.029161777761247423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,5120,0.028765334023369685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,7168,0.27923377354939777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,51200,1.779137717352973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,256,0.04742044541570875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,51200,3.4448212517632375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,512,0.04298488961325752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,1024,0.005721777677536011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,128,0.003759110967318217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,12288,0.22202489111158583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,12288,0.021204445097181533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,4096,0.12099733617570664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,512,0.008560889297061497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,1024,0.028728001647525366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,6144,0.14753511216905382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,3584,0.02404800057411194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,256,0.007913777397738563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,32,0.014030221435758801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,7168,0.04081599911053976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,512,0.0425911115275489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,3584,0.04632000128428141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,512,0.007873777714040544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,3072,0.03422755665249295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,7168,0.14681866433885363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,7168,0.056200888421800405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,32,0.08763288789325291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,3072,0.07063021924760607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,16384,0.6619368659125434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,512,0.005521777603361342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,2560,0.18813244501749674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,3072,0.021733333667119343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,3584,0.008009778128729926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,4096,0.012870222330093384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,7168,0.016533333394262526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,65536,0.23646754688686797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,1536,0.006667555620272954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,7168,0.11272888713412815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,51200,0.055361777544021606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,6144,0.023471110396915015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,65536,1.8930994669596355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,7168,0.28831556108262807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,12288,1.5669955147637262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,2048,0.04036622246106466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,10240,1.4461128446790907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,6144,0.5993768903944228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,512,0.004001777619123459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,7168,0.2405244509379069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,1536,0.011306666665607028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,768,0.03645244571897719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,65536,1.2728586196899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,3072,0.0679617789056566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,7168,0.043041778935326465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,1024,0.025786666406525507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,65536,0.6798951360914441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,5120,0.07041866911782159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,4096,0.03088444471359253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,8192,0.014574221438831754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,8192,0.1974817779329088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,128,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,12288,0.03011111087269253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,3072,2.030630323621962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,7168,0.1602204508251614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,768,0.03277511066860623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,65536,0.1106719970703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,2048,0.00591555568906996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,7168,0.06339289082421197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,768,0.004956444518433677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,4096,0.08588355779647827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,1536,0.011025778121418424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,1024,0.18344622188144258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,6144,0.017795556121402316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,1024,0.005345777918895085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,7168,0.04359377755059136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,512,0.15804089440239802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,512,0.006582222051090664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,8192,0.03309511144955953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,3072,0.9024444156222873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,512,0.006746666712893381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,16384,0.02902488907178243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,51200,0.066557334529029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,768,0.014641778336630927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,3072,0.008556444611814287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,8192,0.016583111551072862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,1024,0.009640889035330879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,16384,0.11728444364335801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,8192,3.9589127434624563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,7168,0.021867555048730638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,2048,0.11897510952419704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,2560,0.00926399976015091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,1536,0.007642666498819987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,128,0.017462222112549674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,2048,0.038134220573637224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,2048,0.012693333129088083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,16384,0.16210399733649358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,1024,0.019848000672128465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,2048,0.007271111011505127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,64,0.00997688869635264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,3072,1.2163395351833768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,128,0.09496888849470352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,64,0.006649777707126405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,51200,0.22237332661946616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,3072,0.029832889636357624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,1536,0.015621332658661736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,12288,0.0488319993019104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,4096,0.10647022061877781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,8192,0.008522666990756989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,6144,0.009977777798970541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,10240,1.0822932985093858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,6144,0.41454577445983887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,2560,0.06642666790220472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,7168,0.07077866792678833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,10240,0.010481778118345471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,512,0.012632889052232107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,1024,0.0070168889231152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,12288,1.441706657409668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,12288,0.020033778415785897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,6144,3.0011698404947915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,3584,0.018168000711335074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,512,0.17460266749064127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,512,0.005522666705979242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,768,0.008403555386596257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,2560,0.007164444360468123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,16384,0.2483688990275065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,768,0.010214221974213919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,4096,0.15371821986304388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,512,0.003834666477309333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,4096,0.01167377746767468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,8192,0.10656799872716267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,768,0.005321777943107817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,32,0.006852444261312485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,16384,0.007744000189834171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,6144,0.051023112403021924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,32,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,16384,0.02158222264713711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,3584,0.008303110798199972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,16384,0.07473155524995592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,32,0.006879111131032308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,2048,0.9679137335883247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,10240,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,256,0.005589333259397083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,51200,0.013697778185208639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,32,0.007997333175606197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,65536,0.3758213255140517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,4096,2.8359991709391275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,128,0.01773422294192844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,64,0.011775111158688864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,3584,0.01633066601223416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,64,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,3072,0.008231110870838165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,2048,0.10788533422682021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,2048,0.017162667380438913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,16384,0.029472000069088403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,51200,1.7952310774061415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,51200,1.058775160047743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,12288,0.028424001402325098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,256,0.004328000048796336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,2048,0.01051288925939136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,512,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,12288,0.07021866904364692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,64,0.015948444604873657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,1536,0.015069334043396844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,3072,0.016193777322769165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,16384,9.87984127468533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,6144,0.08551644616656834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,2560,0.01515111161602868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,6144,0.016417778200573392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,6144,0.039325333303875394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,128,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,6144,0.1277057727177938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,64,0.004507555729813046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,512,0.01864622202184465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,10240,0.0468426677915785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,2048,0.005305777821275923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,8192,0.31382399135165745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,512,0.005789333333571752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,4096,6.761216057671441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,32,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,16384,0.021223111285103693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,768,0.042733334832721286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,1536,0.03636088967323303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,5120,0.4836275312635634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,7168,0.03299466768900553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,10240,0.2841208775838216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,12288,0.06388266881306966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,5120,0.02121777832508087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,51200,0.055797331862979464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,3072,0.008960888617568547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,256,0.003232888877391815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,6144,0.07511911127302381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,512,0.023573334018389385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,12288,0.38115912013583714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,5120,0.8259529007805718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,51200,0.5584675470987955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,4096,0.10089333189858331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,256,0.013007111019558378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,1536,0.025702221526039973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,5120,0.009376000199053023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,3072,0.011362666885058085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,7168,0.1961982250213623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,5120,0.05883644686804878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,8192,0.006176888942718506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,3072,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,12288,0.41820356580946183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,2048,0.2058124409781562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,1024,0.10464444425370957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,16384,0.10505688852734035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,65536,0.12474844190809463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,3072,0.1631457805633545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,7168,12.752570258246529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,768,0.0074879998962084455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,2560,0.007886221839321984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,65536,1.428550190395779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,65536,0.0251217782497406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,32,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,51200,0.06302489174736871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,8192,0.5923831197950575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,2048,0.8796470959981283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,4096,0.03367822368939718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,10240,2.24052980211046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,256,0.02583911187118954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,51200,1.337142202589247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,5120,0.36147112316555446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,256,0.0038115556041399636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,12288,0.04222933451334635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,64,0.005690666536490123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,3072,0.014666666587193808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,512,0.005040888984998067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,64,0.005696889013051987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,5120,0.012502222425407834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,3584,0.11518666479322646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,6144,0.01740533279048072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,4096,0.07767733600404528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,3072,0.012827555338541666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,3072,0.010258666343159145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,2048,0.10350844595167373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,768,0.04908000098334419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,768,0.020399111840460036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,1536,0.007087111473083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,64,0.003104000041882197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,12288,0.4882666799757216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,8192,0.44759109285142684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,2560,0.010198222266303169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,3072,0.0558240016301473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,128,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,6144,0.013020444247457715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,12288,0.03580088747872247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,12288,0.11475911405351426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,2048,0.007672000262472365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,5120,0.012416000167528788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,3072,0.010087110930018956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,8192,0.02437688906987508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,6144,0.01774666706720988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,51200,0.40508267614576554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,1536,0.006702222343948152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,4096,0.012861332959598966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,2048,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,12288,0.008991111483838823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,8192,0.014578666951921252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,16384,0.2345600128173828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,65536,0.21350933445824516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,4096,0.008647110727098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,6144,0.1754337814119127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,512,0.004252444538805219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,8192,0.006145777801672618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,7168,0.08464444345898098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,768,0.003966222206751506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,10240,0.09511378076341416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,3072,0.09312711159388225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,10240,0.2669377856784397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,512,0.0033031110134389666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,4096,0.02787289023399353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,51200,0.0687333345413208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,512,0.008623999853928884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,32,0.02557955516709222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,256,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,2048,0.011360888679822287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,256,0.010648888846238455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,256,0.00498488876554701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,128,0.004943999979231092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,8192,0.05054044392373827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,2048,0.03644799855020311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,6144,0.020432000358899433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,51200,0.13619911670684814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,6144,0.5175342029995388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,51200,0.4226035541958279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,1024,0.011184000306659274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,2560,0.009096889032257928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,1536,0.007732444339328342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,4096,0.00830488900343577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,10240,0.23373156123691133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,1024,0.2289431095123291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,2560,0.014053333136770459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,65536,0.7074089050292969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,65536,16.581644694010418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,16384,1.4244293636745875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,7168,0.07447644737031725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,768,0.020195555355813768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,768,0.006614222294754452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,32,0.04365066687266032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,12288,0.5576693216959635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,10240,0.04052799940109253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,3584,0.010629333555698395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,10240,0.025183111429214478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,65536,0.10720444387859768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,2048,0.007354666789372762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,65536,0.08375555276870728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,1024,0.004249777644872665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,7168,0.05415199862586128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,512,0.005038222091065513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,768,0.012349333200189801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,10240,0.035544888840781316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,3072,0.13248088624742296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,3072,0.007506666911972894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,2560,0.016495111915800307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,16384,0.2372871239980062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,16384,0.018596443865034316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,64,0.00272088891102208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,5120,0.009770666559537252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,8192,0.014170666535695394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,12288,10.352525499131945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,8192,0.7437200016445584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,8192,0.0616497794787089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,7168,0.04367644588152567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,5120,0.014730667074521383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,3584,0.029014223151736792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,7168,0.12338222397698297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,512,0.006272000157170826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,128,0.014017777310477363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,65536,0.10722221930821736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,3072,0.11829333835177952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,3584,1.135372479756673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,1024,0.0443155566851298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,128,0.007487110793590546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,2048,0.007128888534175024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,256,0.01797777745458815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,12288,0.05445955528153313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,5120,0.17358044783274332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,12288,0.2195377747217814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,7168,0.011267555256684622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,1024,0.005483555710977978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,128,0.015198222464985318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,1536,0.0047288888858424295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,2560,0.00792266676823298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,2048,0.006049777898523543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,7168,0.02277155551645491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,512,0.09479022026062012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,1536,1.3819928699069555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,10240,0.019528888993793063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,2048,0.009544889132181803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,32,0.0036880001425743103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,768,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,512,0.006767999794748094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,5120,0.026201777988009985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,8192,0.05995821952819824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,512,0.009514666265911525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,2560,0.01686755485004849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,2048,0.006100444330109491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,64,0.024352888266245525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,768,0.022075555390781824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,32,0.01202755586968528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,128,0.009668444593747457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,3072,0.008055999875068665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,4096,1.0450888739691842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,10240,0.8023031022813586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,51200,0.06683111190795898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,65536,0.09304177761077881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,64,0.025787555509143408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,3072,0.010451555252075195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,3072,0.01882222294807434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,12288,0.13285333580440944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,512,0.017256889078352187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,51200,5.163755628797743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,3584,0.012310222619109683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,2560,0.011733333269755045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,7168,0.0950160026550293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,51200,0.8317786852518717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,7168,0.06444888644748263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,1024,0.049829334020614624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,12288,0.019738665885395475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,768,0.0695742236243354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,32,0.005123555660247803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,8192,0.013796444568369122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,16384,0.07571644253200956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,10240,0.02591733303334978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,8192,0.17704444461398652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,1536,0.014771555860837301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,128,0.011112000379297467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,5120,0.02385066615210639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,4096,0.035989334185918175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,1536,0.013642667068375481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,7168,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,3072,0.014815110299322339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,3584,0.052116443713506065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,12288,3.1311634911431203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,4096,0.009648000200589498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,7168,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,2048,0.09531111187405056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,768,0.007296000089910295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,6144,0.7009155485365125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,256,0.006751110984219446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,2048,0.11760532855987549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,5120,0.027355555031034682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,1024,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,128,0.016557332542207506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,2048,0.01291911138428582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,3584,0.0921048919359843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,12288,0.0292577776643965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,1024,0.004081777814361784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,768,0.022662222385406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,5120,0.12779111332363552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,1024,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,10240,0.4511431058247884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,10240,0.059465779198540576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,64,0.005225777626037598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,768,0.003535110917356279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,512,0.006846222198671765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,51200,1.1020178265041776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,2048,0.007831999825106727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,2048,0.11751200093163384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,10240,0.015878222054905362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,65536,0.07781155241860284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,4096,0.010564444793595208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,10240,0.03456177645259433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,512,0.005446222093370225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,1024,0.005711110929648082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,8192,0.006421333385838403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,10240,0.0357528891828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,3584,0.024686222275098164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,512,0.007522666619883643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,3584,0.016255999604860943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,8192,0.2870231204562717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,768,0.007448888487286038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,256,0.017952000101407368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,768,0.018407111366589863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,8192,0.2646346622043186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,5120,0.011757333245542316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,4096,0.043720000320010714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,6144,0.04307555490069919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,10240,4.6367064581976996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,65536,0.1955404414070977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,10240,0.07681333356433444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,65536,0.06623644298977323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,1024,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,16384,0.16279555691613093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,768,0.004642666627963384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,1024,0.013879111243618859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,3584,0.012311999996503195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,2048,0.04064888755480448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,7168,0.2219324509302775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,8192,0.031677332189348005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,65536,1.0619369082980685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,1536,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,128,0.014327999618318347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,3072,0.014545778433481852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,3584,0.10809599690967137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,7168,0.012621333201726278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,4096,0.03928088810708787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,32,0.008557333714432186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,7168,0.06896977954440646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,1024,0.0088746663596895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,65536,0.18854310777452257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,5120,0.03170844581392076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,3584,0.06726133161120944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,3584,0.01316177762216992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,3072,0.023952000670962866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,128,0.010050666828950247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,51200,6.850115458170573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,32,0.010082667072614035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,8192,0.13780889246198866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,2560,0.03992444607946608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,2048,0.014826666977670459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,7168,0.06274044513702393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,16384,0.23664977815416124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,16384,0.01787644459141625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,3584,0.06158755885230171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,10240,0.015295111470752291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,16384,0.03294577863481309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,10240,3.376795450846354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,1024,0.006635555376609166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,16384,2.4728836483425565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,4096,0.0708106689982944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,2560,0.06879378027386136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,65536,0.6876364284091525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,768,0.011560888753996955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,5120,0.01808177762561374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,7168,0.22860532336764863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,10240,0.034882667991850115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,7168,0.019458666443824768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,3584,0.015523556205961438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,256,0.0028355556229750314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,3584,1.1183226903279622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,768,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,128,0.006735111276308696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,5120,0.02051199972629547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,5120,0.013725333743625216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,51200,0.058535999721950956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,3072,0.009291555318567488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,1536,0.1615999937057495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,7168,0.015440000428093804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,1536,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,51200,0.28849511676364475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,51200,0.10035644637213813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,4096,0.008229333493444655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,1024,0.021617778473430212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,16384,0.026895999908447266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,768,0.02770044406255086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,128,0.0027457777824666765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,2048,0.00756888919406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,7168,0.01457244488928053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,51200,0.3829893271128337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,51200,0.5030941963195801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,10240,0.08991377883487278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,16384,0.04401333464516533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,7168,0.1161208947499593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,12288,0.17706667052374947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,4096,0.012267555627557965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,16384,80.83972846137154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,2560,0.05288533369700114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,32,0.0026195556339290407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,7168,0.018007111218240526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,1024,0.04479111234347025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,51200,0.3693591223822699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,12288,0.30957333246866864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,128,0.1151768896314833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,2048,0.012061333490742577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,51200,0.058916442924075656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,2048,0.019874667127927143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,6144,0.024128889044125874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,51200,0.21099021699693468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,51200,3.2456107669406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,10240,0.27571288744608563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,51200,2.3036524454752603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,5120,0.2554017702738444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,12288,0.5055466757880317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,512,0.0038453332251972626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,768,0.00868266655339135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,10240,0.034154666794670954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,51200,0.18800800376468232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,32,0.014138667119873894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,51200,0.5240809122721354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,1536,0.01293333371480306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,8192,0.0469395551416609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,7168,0.14068088266584608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,2560,0.18618310822380912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,1536,0.009504888620641496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,10240,0.5152453316582574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,1024,0.014370666609870063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,32,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,128,0.2546488973829481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,8192,0.012253333297040729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,51200,0.1182382239235772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,51200,0.14511466026306152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,12288,0.36128444141811794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,6144,0.013159111142158508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,12288,0.07291377915276422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,12288,0.11036088731553818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,12288,0.1128115521536933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,3072,0.012958221965365939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,7168,0.21779378255208334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,4096,0.14446311526828343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,3584,0.010182222558392419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,3072,0.015060444672902426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,2048,0.005511999958091312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,3072,0.10521511236826579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,6144,1.6816346910264757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,5120,0.03921688927544488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,51200,0.08246133062574598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,128,0.22418400976392958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,768,0.006664888726340399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,3584,0.012002666791280111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,2048,0.016937777400016785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,3584,0.058824002742767334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,128,0.006710222197903528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,12288,0.6051697731018066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,1536,0.008772444393899705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,1024,0.01366666704416275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,10240,0.029156443145540025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,768,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,5120,0.7790933185153537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,8192,0.03949688871701559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,1024,0.024336000283559162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,12288,0.02015555567211575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,128,0.008728889127572378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,3584,0.014757333530320061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,3072,0.009528888596428765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,5120,0.010507555471526252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,3072,0.10025955571068658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,3584,0.015634665886561077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,10240,0.11345867315928142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,128,0.004113777644104428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,7168,0.10778844356536865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,2560,0.08954933616850112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,3584,0.12441421879662408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,16384,0.28630044725206166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,10240,0.4726684358384874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,256,0.008990222381220924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,64,0.0030986666679382324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,512,0.01014666673209932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,3584,1.0213626225789387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,65536,0.2047884464263916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,3072,0.018188445104493033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,1536,0.004729777574539185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,512,0.0034088889757792154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,3584,0.05751022365358141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,64,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,6144,0.25332623057895237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,6144,0.025582222474945918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,32,0.050501333342658154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,7168,0.08317955334981282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,10240,0.023507555325826008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,16384,1.9432754516601562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,2048,0.013807111316257052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,2048,0.09212266736560398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,12288,0.02742133372359806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,1536,0.005580444302823808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,64,0.026962666047943964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,7168,0.08305688699086507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,51200,0.0947342250082228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,32,0.003600888782077365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,16384,0.07251822286181979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,12288,0.0177182224061754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,2048,0.01644533375898997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,256,0.16273955504099527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,7168,0.018326222896575928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,51200,0.38551288180881077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,10240,0.08022933536105685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,12288,0.014408888088332282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,12288,0.022847111026446026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,5120,0.015524443652894763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,8192,0.04998666710323758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,12288,0.035107556316587664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,128,0.0031582222630580268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,768,0.02071199980047014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,7168,0.03195022212134467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,3072,0.009895111123720804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,128,0.010439111126793755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,51200,0.14967555469936794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,128,0.004379555583000183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,65536,0.6058879958258735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,65536,3.116547478569878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,2560,0.02344977855682373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,1536,0.06106577979193794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,768,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,4096,0.13369511233435735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,4096,0.04699288805325826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,5120,0.027273777458402846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,768,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,7168,0.1630222267574734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,768,0.008296888735559251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,512,0.04830044507980347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,5120,0.044870220952563815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,2560,0.4068453311920166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,10240,0.04826044374042087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,12288,0.045956446064843066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,16384,0.4129946761661106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,16384,0.022180444664425317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,4096,0.13692355155944824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,12288,0.07939466502931383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,2048,0.007060444189442529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,65536,0.2142559952206082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,65536,0.28373244073655873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,64,0.003371555565132035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,3584,0.01367911116944419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,128,0.011254222028785281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,64,0.010562666588359408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,1536,0.02679466704527537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,3072,0.03232622146606445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,10240,0.025003555748197768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,1024,0.006671999891599019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,4096,0.039891554249657526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,64,0.006421333385838403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,256,0.009887110855844285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,3584,0.09205689032872517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,64,0.03511555658446418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,2048,0.02033955521053738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,10240,0.006768888897365994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,1536,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,8192,0.11101244555579291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,512,0.007361777954631382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,10240,0.9444755978054471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,768,0.05000266763899061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,6144,0.009580444130632613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,3072,0.009750222166379293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,1536,0.008274666965007782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,3584,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,6144,0.017584888471497428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,2560,0.01571644511487749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,12288,3.0934755537245007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,768,0.3161697652604845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,2048,0.009868444667922126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,12288,0.439997328652276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,32,0.01054222219520145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,2048,0.029321776496039495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,2048,0.02311822275320689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,256,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,256,0.008762666748629676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,2560,0.016725333200560678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,6144,0.5956062210930718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,51200,0.38983466890123153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,7168,0.13911733362409803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,4096,0.2979831165737576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,1024,0.06876355409622192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,6144,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,6144,0.006807111203670502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,64,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,8192,0.01606755620903439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,6144,0.04695644312434726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,256,0.012728888955381183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,2048,0.060257779227362744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,8192,0.18238488833109537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,7168,0.03522933191723294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,3072,0.026485333840052288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,8192,0.012640000217490725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,6144,0.013909333282046847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,512,0.011731555892361535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,7168,0.016664889123704698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,16384,6.660188462999131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,4096,0.01389600005414751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,2560,0.009558222360081142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,1024,0.018375110295083787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,1024,0.005818666683302985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,6144,0.27395643128289116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,1024,0.029222223493787978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,256,0.00319377767542998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,6144,0.01460000044769711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,2048,0.04226488868395487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,4096,0.030992888742023047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,64,0.003212444484233856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,6144,0.6894284354315864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,12288,0.023989333046807185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,256,0.007466666400432587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,4096,0.09016266796323989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,6144,0.01770400007565816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,3584,0.016155555844306946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,10240,0.020280000236299302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,51200,0.7718115382724338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,4096,0.8380968835618762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,16384,0.2251208888159858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,2560,0.2498391204410129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,1536,0.045736889044443764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,32,0.1837306684917874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,5120,0.02049955560101403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,5120,0.012771555946932899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,768,0.0034017778105205963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,1024,0.02488533324665493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,7168,0.028428445259730022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,6144,0.34242934650844997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,1024,0.005608888963858287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,128,0.061666667461395264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,2048,0.016378665963808697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,64,0.0025555555605226094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,4096,0.02051999999417199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,16384,0.022112000319692824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,65536,0.15105422337849936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,65536,0.3518328931596544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,5120,0.057714667585161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,2048,0.07502044571770562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,10240,0.09833066993289524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,32,0.015468445089128284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,12288,0.4321182303958469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,256,0.004967111266321606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,512,0.003607999947335985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,16384,0.021781333618693884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,16384,0.04593777656555176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,3072,0.04706133405367533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,12288,0.047188444270028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,768,0.005102222164471944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,4096,0.020921778347757127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,16384,0.028186665640936956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,8192,0.14266488287183973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,512,0.003528000166018804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,7168,0.02185955478085412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,5120,0.08547111352284749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,2560,0.03884444302982754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,51200,0.987693362765842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,8192,0.03856177793608771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,512,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,3072,0.07905066675610013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,512,0.005568000177542369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,512,0.03898755709330241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,6144,0.04983466532495287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,16384,0.019016888406541612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,6144,0.6156631045871311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,16384,0.023912888434198167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,768,0.01791822248035007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,10240,0.023864888482623633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,2048,0.008755555583371056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,2048,0.07968978087107341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,128,0.01052266690466139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,7168,2.434167014227973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,51200,0.2735359933641222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,2560,0.023961777488390606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,768,0.02862311071819729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,5120,0.028997331857681274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,10240,0.023016888234350417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,256,0.006812444577614467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,51200,0.4159395429823134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,7168,0.2780168851216634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,1024,0.003899555653333664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,6144,0.01440355512830946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,1024,0.006312888943486744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,128,0.1533084445529514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,6144,0.03303822212749057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,64,0.007228444847795699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,2048,0.01604711181587643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,768,0.004804444395833545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,1024,0.005675555517276128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,1536,0.06703733073340522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,2560,0.11761155393388535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,256,0.004753777964247597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,4096,0.04091466797722711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,2048,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,7168,0.7372515466478137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,4096,0.16224000189039442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,3584,0.10744088888168335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,768,0.15597689151763916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,51200,0.2150204446580675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,768,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,10240,0.04201155569818285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,768,0.00982488857375251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,12288,0.04640444450908237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,4096,0.017438222964604694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,10240,0.3032746579911974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,128,0.0028560000161329904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,2560,0.8037724494934082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,7168,0.015933333171738517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,12288,0.2225022315979004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,51200,0.06705600023269653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,65536,1.9292827182345922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,10240,0.14595111211140951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,1536,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,1024,0.007575111256705389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,2560,0.007859555383523306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,2048,0.012410666379663678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,5120,0.20616888999938965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,7168,0.03295911020702786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,128,0.003428444473279847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,16384,0.04829422301716275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,8192,0.11534310711754693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,12288,0.029127111037572224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,16384,0.4619510968526204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,512,0.0052622221410274506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,128,0.004804444395833545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,32,0.03636533353063796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,10240,0.02036977807680766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,1024,0.007516444557242923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,5120,0.012851555314328937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,7168,0.01962311069170634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,51200,0.0696257750193278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,5120,0.011094222466150919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,512,0.011787555283970304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,3072,0.03963911202218797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,7168,0.025905778010686237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,512,0.016951110627916124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,1024,0.009628444082207149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,2048,0.0239991115199195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,8192,0.011944889194435544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,51200,0.053074667851130165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,4096,0.01707555519209968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,4096,0.025247111916542053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,16384,0.05612444215350681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,65536,0.07690755526224773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,3072,0.007654222349325816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,2048,0.06941066847907172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,12288,0.1479226615693834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,7168,0.012714666624863943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,6144,0.056088891294267446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,4096,0.03192266821861267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,1024,0.008015111088752747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,2560,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,768,0.008220444122950235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,3584,0.03212799959712558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,1024,0.00923733330435223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,7168,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,12288,0.03673155440224542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,65536,0.0672960016462538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,2560,0.007088888850477006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,64,0.002830222249031067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,256,0.013453333742088742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,51200,0.35839909977383083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,12288,0.03283910950024923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,8192,0.02533333334657881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,12288,0.006189333481921091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,5120,0.04160355528195699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,12288,0.061646223068237305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,2560,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,512,0.015885333220163982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,768,0.024155555499924555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,512,0.005783111270931032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,768,0.007677333222495184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,12288,5.70304192437066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,10240,0.03055022160212199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,1536,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,128,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,2560,0.008388444781303406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,3072,0.015240000353919135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,16384,0.473599116007487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,256,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,32,0.2069422271516588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,7168,0.12465688917371963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,12288,0.6525688701205784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,64,0.02401955591307746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,2048,0.008917333351241218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,16384,0.12791200478871664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,7168,0.0789511137538486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,12288,0.06453422043058607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,1536,0.03156088789304098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,64,0.009708444277445475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,51200,0.22935112317403158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,4096,0.13478933440314397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,512,0.087190224064721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,64,0.00849599970711602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,256,0.0076453329788313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,3072,0.016726222303178575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,128,0.0052168890833854675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,5120,0.03787466552522447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,6144,0.040234665075937905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,32,0.023709333605236475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,10240,0.01532444523440467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,5120,2.087010701497396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,65536,0.17561066150665283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,5120,0.008047111332416534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,10240,0.05783910883797539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,64,0.003253333270549774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,16384,0.0210515558719635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,3584,0.008385777473449707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,512,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,1024,0.004584889031118817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,7168,2.262242635091146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,7168,0.0874648888905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,128,0.034640000926123724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,32,0.03565866748491923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,2048,0.010444444086816577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,6144,0.014013333453072442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,1024,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,12288,0.030327998929553564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,7168,0.03181333343187968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,16384,1.0302826563517253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,256,0.044237332211600415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,12288,0.019909333851602342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,2560,0.02611200014750163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,1536,0.006073777874310811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,64,0.004266666869322459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,128,0.018138666947682697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,4096,0.04811822374661764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,3072,0.008803555534945594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,256,0.0032266666077905228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,10240,0.01752000053723653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,2560,0.007126222054163615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,3072,0.1164142158296373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,3072,0.019863999552196927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,16384,0.03908888830078973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,51200,0.1652053329679701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,8192,0.019187556372748483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,5120,0.09327377875645955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,1024,0.12433511681026882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,12288,0.10430310832129584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,512,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,51200,0.019677332705921598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,3584,0.10876088672214085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,1536,0.006582222051090664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,51200,0.20362844732072619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,51200,0.5574426651000977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,12288,1.4645813835991754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,4096,0.01203555530971951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,12288,0.0248026665714052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,1024,0.016319110989570618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,128,0.007387555307812161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,16384,0.8593733575608996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,768,0.017717333303557504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,8192,0.031104889180925157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,32,0.00888533310757743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,7168,0.06371733215120104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,128,0.0028213332924577924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,8192,0.01998666591114468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,1536,0.05531733234723409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,5120,0.156131558948093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,2048,0.8945715162489148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,4096,0.03923022084765964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,12288,0.030652443567911785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,1536,0.03399466805987888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,768,0.0058400001790788435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,51200,0.24239023526509604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,4096,0.05694488684336344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,1536,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,5120,0.16546666622161865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,12288,0.030488000975714788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,128,0.006048000107208888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,16384,0.062260442309909396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,8192,0.2623271147410075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,128,0.009452444811662039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,4096,0.01071466671095954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,256,0.06147378020816379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,64,0.010135110881593492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,2560,0.04984977841377258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,256,0.032646222247017756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,8192,0.03397688931889004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,7168,0.012190221912331052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,5120,0.017866666118303936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,32,0.005018666800525454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,768,0.010750222537252637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,10240,0.11362044016520183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,128,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,768,1.1717253790961373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,65536,1.651439136928982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,8192,0.009663999908500248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,256,0.01609777741962009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,5120,0.017302221722073026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,2048,0.01698755555682712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,512,0.005479111025730769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,51200,8.399046156141493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,768,0.005299555758635203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,10240,0.039263112677468195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,4096,0.2854364448123508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,512,0.004193777839342753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,2560,0.011876444849703046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,3584,0.03978755407863193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,768,0.008008889026112026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,16384,0.06234755780961779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,64,0.02155911094612545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,3072,0.00962400022480223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,2560,0.007792889244026608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,8192,0.015793777174419828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,8192,0.016150222884284127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,64,0.0025759999536805684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,4096,0.11862399843004014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,65536,0.7366275787353516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,8192,0.017596445149845548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,1536,0.006875555548402999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,10240,0.4760764439900716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,64,0.013878222141000958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,768,0.037552889850404524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,10240,0.10235999690161811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,2560,0.07248266537984212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,2560,0.009220444493823582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,65536,4.18074205186632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,3072,0.016499555773205228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,12288,0.06027377976311577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,128,0.008968888885445064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,4096,0.3853946791754828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,1536,0.02126311096880171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,3072,0.016766221986876596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,2560,0.00976533359951443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,51200,0.4288017749786377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,1536,0.06571111414167616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,7168,0.078584889570872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,8192,0.1775439977645874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,128,0.02068800065252516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,64,0.01146666705608368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,2560,0.012595555848545499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,3072,0.00816533300611708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,3584,0.10262044270833333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,5120,0.2337413363986545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,768,0.015630222029156156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,1536,0.010311110979980892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,3072,0.010950222611427307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,7168,0.011713777979214987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,1536,0.01258666647805108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,1024,0.04004888733228048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,7168,0.040694223509894475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,7168,0.24688977665371367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,1536,0.015964445140626695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,8192,0.26325866911146373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,8192,0.6637262238396539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,7168,0.06443111101786296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,64,0.03737955623202854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,768,0.015503111812803479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,1536,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,12288,1.0637608634101021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,256,0.011628444823953839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,5120,0.014780445231331719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,5120,0.06215555800331963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,8192,0.015676443775494892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,4096,0.01127466642194324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,7168,0.02918044394916958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,7168,0.04221066501405504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,16384,0.024122666981485155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,7168,0.02791644467247857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,6144,0.01664444473054674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,512,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,5120,0.011170667078759937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,4096,0.021407999926143225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,3072,0.3619653383890788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,51200,4.886336008707683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,10240,0.2131333351135254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,128,0.005286222116814719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,3584,0.997977786593967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,2048,0.07023911343680488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,32,0.00645955569214291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,1536,0.007162666983074612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,512,0.012877333495352002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,12288,0.049103111028671265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,1536,0.007808888951937358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,10240,0.7590515348646375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,12288,0.057342224650912814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,5120,0.7984204292297363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,64,0.009316444396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,5120,0.10089688830905491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,1536,0.037031110790040754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,1536,0.007861333588759104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,1024,0.005320888840489917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,1024,0.07955555783377753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,8192,0.09789955615997314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,2560,0.006823110911581252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,10240,0.023367111881573994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,128,0.010545777777830759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,1024,0.006128888991143968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,12288,0.012515555653307172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,5120,0.09280533260769314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,3584,0.020431111256281536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,256,0.018704889549149405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,51200,0.03370044297642178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,7168,0.01237777786122428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,32,0.005120888766315249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,2048,0.01447111037042406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,7168,0.019479999939600628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,51200,0.3574204444885254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,3072,0.02752711044417487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,1024,0.04816710948944092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,2048,0.1973439984851413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,8192,0.16606577237447104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,4096,0.05980622106128269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,2560,0.02286400066481696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,16384,0.39185155762566465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,7168,0.34166044659084743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,16384,0.15679200490315756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,1024,0.01273155543539259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,128,0.003240888938307762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,1024,0.5883635414971246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,4096,0.029901332325405542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,32,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,256,0.01534933348496755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,16384,1.828080071343316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,1024,0.009473777479595607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,1536,0.03415111038419936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,7168,0.05240711238649157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,7168,0.0878702203432719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,3072,0.07418577538596259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,16384,0.1405280033747355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,51200,0.24428176879882812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,3072,0.006548444430033366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,768,0.09803377919726902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,16384,0.47969865798950195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,64,0.0031271111220121384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,4096,0.01982933282852173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,6144,0.2405466768476698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,64,0.05968888600667318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,3584,0.10161866744359334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,512,0.008608000146018134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,256,0.006729777902364731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,2560,0.04429511229197184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,128,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,4096,0.07868355512619019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,4096,0.03881599836879306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,10240,0.01367288910680347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,12288,0.10397777954737346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,64,0.0026657777941889236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,6144,0.07564533419079252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,512,0.02603111167748769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,512,0.037478221787346735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,3072,0.08839644326104058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,51200,0.7014773156907824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,1536,0.028314666615592107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,5120,0.009359999663299983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,1536,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,2048,0.005110222018427319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,512,0.018343110879262287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,2560,0.01666755477587382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,1536,0.006187555690606435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,1536,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,2560,0.10655377970801459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,1024,0.004201777693298128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,2560,0.010258666343159145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,1536,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,10240,0.28526756498548717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,10240,0.02581688927279578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,16384,0.15093333191341823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,6144,0.0064791109826829694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,512,0.04856266578038534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,4096,0.013529777526855469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,512,0.0030497778207063675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,1024,0.011504888534545898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,5120,0.012146666646003723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,16384,0.4273564550611708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,6144,0.02723377777470483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,768,0.006933333145247565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,512,0.04332355658213297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,65536,4.959282769097222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,6144,0.021695110532972548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,768,0.006882666713661618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,64,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,65536,0.9995501836140951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,7168,0.03016177813212077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,51200,0.24022934171888563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,8192,0.03128444486194187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,4096,0.020023110840055678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,32,0.0027457777824666765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,65536,3.3905520968967013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,2048,0.009090666969617208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,768,0.012608889076444836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,2048,0.029935999049080744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,6144,0.028869334194395278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,12288,0.026308443811204698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,16384,0.37483734554714626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,32,0.0034097778714365433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,3584,0.009001777403884465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,65536,0.023491554790072974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,64,0.02055377761522929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,4096,0.048560887575149536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,7168,0.3949706554412842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,256,0.01706133286158244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,512,0.008072000410821702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,3584,0.047247999244266085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,3584,0.015757333901193406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,10240,0.3071217801835802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,256,0.005277333160241445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,8192,0.036672890186309814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,3072,0.026054221722814772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,51200,0.05330666568544176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,32,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,2560,0.11926755640241836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,16384,0.11382133430904812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,2048,0.037647111548317805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,65536,0.4691440264383952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,16384,0.021744888689782884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,512,0.023225777679019507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,2048,0.02211644417709774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,2048,0.013596444494194455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,10240,0.5773066414727105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,1024,0.02313333253065745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,256,0.017557332913080852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,12288,0.028917332490285236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,128,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,4096,0.062301331096225314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,5120,0.19434577888912627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,4096,0.006257777826653586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,8192,0.05350222190221151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,16384,0.10417066680060492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,2048,0.07223021984100342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,32,0.0027555556346972785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,65536,0.48588890499538845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,768,0.0047315553658538395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,4096,0.012529777983824411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,3584,0.010132444401582083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,1536,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,512,0.14211022853851318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,10240,0.017461333009931777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,6144,0.31097422705756295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,5120,0.029061334000693426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,65536,0.65374755859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,128,0.0038657776183552216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,8192,0.012554667062229581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,64,0.0052986666560173035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,128,0.024195555183622573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,768,0.013000888956917657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,128,0.042688889635933765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,1536,0.00618577789929178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,128,0.00517155561182234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,6144,0.34054843584696454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,16384,0.014869333969222175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,51200,0.0948186649216546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,7168,0.007176000210973952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,3584,0.05676711267895169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,4096,0.022672888305452134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,3072,0.3004079924689399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,64,0.002630222174856398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,51200,0.3632826540205214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,256,0.009999111294746399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,2048,0.007090667055712805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,512,0.004401777767472797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,3584,0.011796444654464722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,8192,0.007976888782448238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,3584,0.015818667080667283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,64,0.005036444299750858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,1024,0.029675556553734675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,128,0.0031991110493739447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,7168,0.011015111373530494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,2560,0.00814400033818351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,256,0.0042071110672420925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,4096,0.011184000306659274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,16384,0.31510843171013725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,4096,0.021390222840838965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,12288,0.019380443625979953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,768,0.010072888599501716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,3584,0.03824177715513442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,8192,0.007234666910436418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,64,0.0027955555253558685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,256,0.0028782222006056043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,8192,5.832767062717014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,8192,0.04483377933502197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,5120,0.05992799997329712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,512,0.006774222271309958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,4096,0.012031111452314587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,512,0.004064889003833135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,256,0.007144889069928064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,6144,0.13954133457607693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,256,0.06089155541525947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,2048,0.01313866674900055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,1536,0.28368088934156627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,512,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,65536,0.012879110872745514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,256,0.003024000053604444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,128,0.21831733650631377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,2048,0.02158311174975501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,65536,0.2835093339284261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,1024,0.6727697584364148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,8192,0.018397332893477548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,51200,0.06239733431074354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,64,0.016771554946899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,64,0.006835555450783835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,2560,0.01802399920092689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,1536,0.0075395554304122925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,1536,0.017998221847746108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,128,0.015223999818166098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,4096,0.009466666314336989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,6144,0.04796888762050205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,3072,0.01351111133893331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,64,0.04318933354483711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,256,0.004849777867396672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,128,0.0031271111220121384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,1536,0.009511111511124505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,1024,0.009289777941173976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,7168,0.4696515401204427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,8192,0.012019555601808759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,3584,0.014508444401952954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,65536,0.14625422159830728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,512,0.003959111041492886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,768,0.014152000347773233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,64,0.0070168889231152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,1024,0.46200980080498594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,16384,0.1725697782304552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,2560,0.030287110143237646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,8192,0.00905866672595342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,1536,0.12024799982706706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,2048,0.026325333449575637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,1536,0.011740444435013665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,1024,0.006204444501135085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,51200,0.0539902216858334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,2560,0.06948710812462701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,65536,0.8638213475545248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,16384,0.6088124381171333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,10240,0.39649687872992617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,8192,0.15059378412034777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,5120,0.009619555539555019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,2048,0.2593759960598416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,2560,0.00831911133395301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,768,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,256,0.006069333189063602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,6144,0.30771732330322266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,256,0.014023999373118082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,3072,0.01924800044960446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,65536,4.410681830512153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,12288,0.12365867031945123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,512,0.012294222083356647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,1024,0.005582222094138463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,51200,8.512319776746962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,4096,0.04600444436073303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,32,0.010195555786291758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,51200,0.8875146441989474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,3072,0.086008886496226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,4096,0.006582222051090664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,51200,1.847699483235677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,512,0.1736053360833062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,12288,0.08312888940175374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,65536,0.22012800640530059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,256,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,768,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,8192,0.1319217814339532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,64,0.005407111098368962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,256,0.07113777928882174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,3584,0.14956177605523005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,16384,0.03331911232736375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,10240,0.08072266976038615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,6144,0.01847288840346866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,3584,0.131059553888109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,51200,0.5337857670254177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,16384,0.12588711579640707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,5120,0.023169777459568445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,2048,0.006585777633719974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,512,0.00908622228437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,7168,0.06263289186689588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,768,0.003628444340493944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,3072,0.014917333920796713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,3072,0.020440000626775954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,512,0.07191733519236247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,12288,0.01290933373901579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,4096,0.006182222316662471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,3072,0.006051555689838197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,2560,0.010149333212110732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,51200,1.3168071111043294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,2560,0.013895110951529609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,7168,0.016064888901180692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,4096,0.014230221509933472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,7168,0.6017893155415853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,256,0.009786666267448002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,256,0.006143110907740063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,256,0.06410311328040229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,10240,0.9838302400377062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,1536,0.2890568839179145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,12288,0.01901155544651879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,128,0.010231110784742568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,51200,0.9421884748670789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,4096,0.01313866674900055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,3584,0.012695999609099494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,12288,0.019874667127927143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,7168,0.023297778434223596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,7168,0.012304888831244575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,128,0.017822222577200997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,1024,0.016198222835858662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,32,0.015004444453451367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,51200,1.1698178185356987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,16384,0.08814222282833523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,3584,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,128,0.005504000104135937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,32,0.021242666575643752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,6144,0.15060710906982422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,1536,0.02144177754720052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,3584,0.01241422196229299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,2048,0.04018488857481215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,128,0.017634666628307767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,12288,0.006210666563775804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,2560,0.008571555217107138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,64,0.009789333575301701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,6144,0.19410844643910727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,7168,0.04739200075467428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,256,0.003463111196955045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,5120,0.04718666606479221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,10240,0.03507377703984579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,512,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,3072,0.017122666041056316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,768,0.009036444127559662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,512,0.060982220702701144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,1536,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,65536,0.03231911195649041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,1536,0.01742755538887448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,3072,0.05138488941722446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,16384,0.04276444514592489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,2560,0.007151111132568783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,2048,0.013268444273206921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,8192,0.1495680014292399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,10240,0.5026639832390679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,65536,0.8033999866909451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,5120,0.019856888386938307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,32,0.1140364408493042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,768,0.008535111116038429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,65536,0.4678453339470758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,512,0.060248891512552895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,65536,4.494543287489149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,2048,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,32,0.006555555595291986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,16384,0.11459822124905056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,6144,0.15559644169277614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,65536,0.3882435427771674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,5120,0.021746666895018682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,768,0.02805866797765096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,3584,0.007622222105662028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,6144,0.08744177553388809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,1024,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,128,0.017521777086787753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,1536,0.0068959999415609576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,10240,0.09798133373260498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,8192,0.46875646379258895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,51200,0.11397243870629205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,512,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,65536,0.9107671313815646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,768,0.005647999958859549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,7168,0.3528239991929796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,8192,2.570516374376085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,10240,0.0602817800309923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,64,0.0032915555768542816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,128,0.006145777801672618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,16384,2.842781278822157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,256,0.033943110042148165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,2560,0.08841066890292698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,7168,0.15803644392225477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,1024,0.00630044440428416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,1536,0.033766223324669734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,6144,0.28249867757161456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,3584,0.1467075612809923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,10240,0.0412986675898234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,128,0.004234666625658671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,5120,0.013171555267439948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,512,0.004402666870090696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,1024,0.004421333471934001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,51200,0.7750897937350802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,1024,0.0058115555180443665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,128,0.010969777901967367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,16384,0.7143600251939562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,2048,0.019111111760139465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,7168,0.09033422337638007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,2048,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,6144,0.11529510551028782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,5120,0.09149955378638373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,6144,0.028068443139394123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,65536,0.909621344672309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,5120,4.953902350531684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,128,0.004032888760169347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,4096,0.035094221433003746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,5120,0.16573066181606716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,512,0.0796568857298957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,7168,0.023660444551044043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,32,0.006215999937719769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,12288,0.21757422553168404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,65536,0.06683111190795898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,3072,0.03139822350607978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,256,0.004320888883537716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,256,0.007500444849332173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,51200,0.21215111679501006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,256,0.005117333183685939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,4096,1.1155662536621094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,4096,0.1257235606511434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,51200,0.3729413350423177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,7168,0.12815822495354548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,3072,0.14411732885572645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,5120,0.015662221444977652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,768,0.01347555551264021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,256,0.0032088889016045462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,128,0.0053386667536364655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,6144,0.06788711415396796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,2560,0.01298577752378252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,8192,0.008960888617568547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,2048,0.036870221296946205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,16384,0.07367377811008029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,7168,0.39429423544141984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,8192,0.060695111751556396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,16384,0.02477866742346022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,32,0.006844444407357111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,3584,0.011384888655609555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,4096,0.011507555842399597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,12288,0.2609768973456489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,128,0.0063386667105886675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,512,0.005520888914664586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,512,0.052319112751219005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,2560,0.006833777659469181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,4096,0.064264886909061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,65536,0.7599324650234647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,768,0.022441777918073866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,128,0.009113777842786577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,65536,0.182060440381368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,3584,0.04174133472972446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,3072,0.11323022842407227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,16384,0.021683555510309007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,10240,0.02905333373281691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,512,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,4096,0.13147377967834473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,6144,0.0162808895111084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,32,0.0058257778485616045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,256,0.005132444616821078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,6144,0.012411555482281579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,64,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,5120,0.014109333356221518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,1536,0.03406577640109592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,1024,0.049885332584381104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,64,0.004918222212129169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,768,0.005670222143332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,5120,0.037626667155159846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,10240,0.1536737812889947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,16384,1.070683585272895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,2048,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,2560,0.18708177407582602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,512,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,12288,0.42412977748446995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,256,0.005663110978073544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,8192,0.09891821940739949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,2048,0.011071110765139261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,3584,0.055789331595102944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,10240,2.047513749864366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,12288,0.035790221558676824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,1024,0.011858666936556498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,51200,11.00701649983724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,2048,0.016229333149062265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,2048,0.36744266086154515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,51200,0.0627688897980584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,128,0.0070675553547011475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,12288,0.020826665891541373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,8192,0.029096888171301946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,128,0.007357333269384172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,51200,0.14366043938530818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,6144,0.049351112710105054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,12288,0.09970221916834514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,10240,0.12649510966406927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,65536,0.1549102200402154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,64,0.005365333209435145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,512,0.004287999951177173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,4096,0.038796444733937584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,1536,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,3584,6.294793870713975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,64,0.009003555609120263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,12288,0.010830221904648675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,64,0.00979555563794242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,65536,0.09509866767459446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,3584,0.009401777552233802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,256,0.0073928890956772705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,4096,0.0236826671494378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,6144,0.551529778374566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,3584,0.009443555441167619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,1024,0.010246222217877706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,4096,0.019370666808552213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,128,0.006610666712125142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,5120,0.012561777399645912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,2048,0.04160977734459771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,64,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,2048,0.005487999982304043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,6144,0.22799465391370985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,256,0.011505777637163797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,128,0.005192888693677055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,1536,0.004953777624501122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,16384,0.0396399994691213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,6144,0.009858667022652095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,12288,0.05708533525466919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,3584,0.00868622213602066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,1536,0.012544000314341651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,65536,0.11249332957797581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,256,0.014304889572991265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,256,0.0041911109454101985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,65536,0.4072222179836697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,65536,0.014923555983437432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,32,0.005501333210203383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,3584,0.01687111126052009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,10240,0.022483555807007685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,1536,0.006551110910044775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,32,0.006783999916579988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,2048,0.04740800129042732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,768,0.004068444586462445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,10240,0.05865066581302219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,256,0.005129777722888523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,51200,0.8657671080695258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,8192,0.04439466529422336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,768,0.053949332899517484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,1024,0.009404444032245213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,2560,0.08033244477377997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,16384,0.007408888803588019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,5120,0.040531555811564125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,768,0.008033778104517195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,2048,0.13638044728173151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,768,0.006158222340875202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,2048,0.01921955578856998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,5120,0.11343733469645183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,7168,0.034127109580569796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,65536,0.2967040008968777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,16384,0.024328889118300542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,6144,0.01146133326821857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,128,0.007324444750944774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,51200,1.7768756018744574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,256,0.008061333662933773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,1536,0.0075839997993575195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,32,0.011815110842386881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,128,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,10240,0.05248622099558512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,128,0.0030497778207063675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,2048,0.10979377561145359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,32,0.08962044450971816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,10240,0.020383111304706998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,7168,0.0692853331565857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,768,0.005030222237110138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,256,0.08609600199593438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,2048,0.010214221974213919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,3072,0.0162462227874332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,6144,0.009672000176376766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,128,0.004548444516128964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,10240,0.03332355618476868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,3584,0.05793866846296522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,6144,0.016531555189026725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,3072,0.00904266701804267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,65536,3.7971411810980906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,4096,0.02114222281508976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,128,0.0029564443975687027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,1024,0.004420444369316101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,64,0.006871999965773688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,16384,0.023503111468421087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,65536,0.6922639740837945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,12288,0.00849333322710461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,5120,0.02906222144762675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,65536,0.08185599909888373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,6144,0.016350222958458793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,128,0.003253333270549774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,256,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,4096,0.013668444421556262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,64,0.010195555786291758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,512,0.21901599566141763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,8192,0.7129484282599555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,12288,0.028817776176664565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,16384,0.028484443823496502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,8192,0.014926221635606555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,2048,0.03783911135461595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,1536,0.10960266987482707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,64,0.15814577208624944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,12288,0.9777599970499674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,6144,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,8192,0.033300442828072443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,6144,0.12596000565422907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,2048,0.01203555530971951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,1536,0.008624888956546783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,1536,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,6144,0.24663734436035156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,51200,4.4299091762966585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,2048,0.009222221871217092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,10240,0.23268177774217394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,2048,0.011284444067213269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,4096,0.02243822150760227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,16384,1.45813783009847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,2560,0.01036266651418474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,256,0.03335111008750068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,8192,0.01707466608948178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,16384,0.03033422099219428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,256,0.014653333359294467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,3584,0.007864889171388414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,8192,0.10728622145122951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,16384,0.07219377491209242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,4096,0.04226933254135979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,3072,0.13941778077019587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,768,0.04976266622543335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,2048,0.013611555927329592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,2048,0.025942222939597234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,32,0.0029973333908451926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,6144,0.010598222414652506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,3584,0.016393777396943834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,8192,0.010556444525718689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,65536,0.32379378212822807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,6144,0.8103084564208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,16384,0.42968000306023496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,5120,0.010645333263609143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,16384,0.428007099363539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,65536,0.0736008882522583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,5120,3.264752917819553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,5120,0.09062400129106309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,768,0.030616889397303265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,51200,0.24014666345384386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,65536,0.9428720474243164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,51200,0.11010933584637111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,65536,0.10150311390558879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,16384,0.9058924780951606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,12288,0.0938746664259169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,512,0.04359289010365804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,10240,0.05993600024117363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,1536,0.11377067036098903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,1024,0.0058808889653947614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,4096,0.2936551041073269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,2048,0.007134222322040134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,8192,0.03418755531311035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,1536,0.04229688975546095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,51200,2.4608063167995877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,1024,0.04811199837260776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,3072,0.011728888584507836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,1536,0.009767110976907942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,1536,0.064337776766883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,3584,0.00906666699382994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,16384,5.329578823513455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,32,0.02130577796035343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,32,0.002962666667169995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,16384,7.018199920654297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,256,0.07324177689022489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,4096,0.044751998451020986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,1536,0.005960000058015187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,1024,0.0228426671690411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,7168,0.03238666719860501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,10240,0.5035884645250108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,7168,0.09229599767261082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,768,0.007033777733643849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,512,0.003914666672547658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,512,0.008966222405433655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,128,0.002973333415057924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,6144,2.4859937032063804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,512,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,2560,0.0547253323925866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,512,0.02014755540423923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,128,0.01260355528857973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,64,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,1024,0.007335111498832703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,3584,0.12661511368221706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,1024,0.08223199844360352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,1536,0.00832266691658232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,1024,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,32,0.005069333232111401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,2048,0.09352799918916489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,1536,0.02066044509410858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,51200,1.1519716050889757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,32,0.005596444424655702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,16384,0.04310133390956455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,1536,0.11323200331793891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,4096,0.09490666786829631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,4096,0.019175110591782462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,1024,0.2380631234910753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,8192,3.102657741970486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,3584,0.02754222187731001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,51200,0.06520977947447035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,5120,0.02793777651256985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,256,0.012433778080675336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,64,0.005119110975000593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,6144,0.01573244399494595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,768,0.01866044435236189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,51200,0.41002488136291504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,3584,0.15345599916246203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,6144,0.2319128778245714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,1024,0.027615999182065327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,65536,8.22101084391276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,2560,0.024326221810446844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,5120,0.007663999994595845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,6144,0.07984089189105563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,5120,0.18500710858239067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,512,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,512,0.02418488926357693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,8192,0.061793777677747935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,1536,0.12170399559868707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,5120,0.006720888945791457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,2560,0.012683555483818054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,1536,0.005997333261701796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,1024,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,768,0.061976889769236244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,512,0.0033857776886887024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,12288,6.080130683051215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,12288,0.022006221943431433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,4096,0.008711111214425828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,12288,0.023899555206298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,5120,0.09112444188859727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,16384,0.01977777812216017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,65536,0.24883998764885795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,10240,0.4542995558844672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,12288,0.028746667835447524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,4096,0.02646933330429925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,65536,0.31964267624749076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,256,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,7168,0.028408888313505385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,51200,0.08617955446243286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,768,0.005054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,10240,0.021751999855041504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,768,0.013799111048380533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,4096,0.1340871122148302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,16384,0.08656444152196248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,256,0.005058666484223472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,10240,0.2864862283070882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,1024,0.03683911098374261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,4096,0.009418666362762451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,6144,0.021293333835071985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,10240,0.3051244417826335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,512,0.03610400027698941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,7168,0.6463653246561686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,256,0.007253333098358578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,16384,0.2062364419301351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,51200,1.1903893152872722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,512,0.008260444634490544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,3584,0.09780000315772162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,768,0.005173333403136996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,64,0.09305866559346516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,768,0.0070186663005087115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,10240,0.2966266738043891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,6144,0.0808853308359782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,256,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,16384,0.1368302239312066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,128,0.0030906666070222855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,1536,0.018481777773963082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,256,0.004854222138722737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,4096,0.1863457759221395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,32,0.03202399942609999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,8192,0.10640533102883233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,2048,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,512,0.24540445539686415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,8192,2.068868425157335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,4096,0.06488177511427137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,65536,0.11754933993021648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,32,0.005227555417352253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,3584,0.12785600291358099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,4096,0.015453333655993143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,128,0.005771555420425203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,3072,0.1263093286090427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,4096,0.006072000082996156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,1024,0.006140444427728653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,2048,0.026357332865397137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,2560,0.08252533276875813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,5120,0.11599555280473496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,2048,0.12924177116817898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,2048,0.012488000094890594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,6144,0.028007109959920246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,3072,0.0081386665503184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,128,0.003044444446762403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,7168,0.011369778050316704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,32,0.005505777895450592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,65536,0.1042284435696072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,32,0.005250666704442766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,512,0.005575110928879843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,65536,0.48592175377739805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,51200,0.038514667087131076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,512,0.006385777973466449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,6144,0.04218488931655884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,5120,0.014107555150985718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,3584,0.028527110815048218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,51200,0.24936445554097494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,512,0.010703999963071613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,8192,0.9301528930664062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,6144,0.03166133496496413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,16384,0.11847910616132949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,65536,5.0148133171929254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,51200,0.2654835648006863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,2048,0.011721777419249216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,128,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,512,0.008561777571837107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,65536,0.011048888994587792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,7168,1.5978497399224176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,5120,0.13090844949086508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,2560,0.06496355268690321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,2560,0.04680888851483663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,12288,0.04960888955328199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,16384,0.02588355541229248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,8192,0.012904889053768583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,512,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,6144,0.021190222766664293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,2560,0.008179555336634317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,65536,5.727457682291667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,32,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,3584,0.04665689004792107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,7168,0.009138666921191746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,512,0.005464000006516774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,12288,0.021864889396561518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,51200,2.098879072401259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,256,0.003896000070704354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,768,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,65536,4.836809794108073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,128,0.008766222331258986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,768,0.006100444330109491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,1024,0.005417777846256892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,6144,0.013178666432698568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,768,0.005706666658322017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,3072,0.557728025648329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,4096,0.22988888952467176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,2048,0.01900800069173177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,128,0.005782222168313132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,1536,0.007222221957312689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,768,0.029003553920321997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,128,0.003927111211750242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,32,0.009871111147933537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,4096,0.008788444101810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,5120,0.04607466525501675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,7168,0.032607999112870954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,5120,0.015662221444977652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,65536,0.18979733520083955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,1536,0.3369493219587538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,65536,1.7173431184556749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,12288,0.6076915529039171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,128,0.0125928885406918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,768,0.09298933214611477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,1024,0.056539555390675865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,32,0.00656355544924736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,3584,0.011750222080283694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,256,0.2588142289055718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,5120,0.007872888611422645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,6144,0.9700436062282987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,10240,0.047705776161617704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,128,0.005953777581453323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,4096,0.07204088899824354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,65536,1.7427084181043837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,2560,0.03380799955791897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,4096,0.014639111028777229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,51200,0.03059200114674038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,16384,0.4372888935936822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,5120,0.016311999824311998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,65536,1.9196408589680989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,64,0.0028008888992998335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,8192,2.06478394402398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,10240,0.016855110724767048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,3584,0.008616000413894653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,16384,0.36173423131306964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,768,0.036899556716283165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,6144,0.03200088938077291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,3072,0.011752000285519494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,768,0.009534222384293875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,5120,0.014306666122542487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,12288,0.6225804222954644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,8192,1.8304710388183594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,2560,0.0240346673462126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,3584,0.36604266696506077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,5120,0.014486221803559197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,8192,0.031714667876561485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,64,0.06840533018112183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,64,0.006752000086837345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,4096,0.08866133292516072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,16384,0.741121768951416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,5120,3.4161963992648654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,3072,0.010275555153687796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,7168,1.689927101135254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,2048,0.018031110366185505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,3072,0.06807555754979451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,128,0.01222222215599484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,32,0.04109422365824381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,64,0.0028693332440323303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,512,0.005737777799367905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,1024,0.05394400159517924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,10240,0.19637778070237902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,2048,0.011063111325105032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,12288,0.13999556170569524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,12288,0.1723111073176066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,10240,0.026545777916908264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,512,0.026400888959566753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,10240,0.01758311192194621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,3072,0.07309333483378093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,16384,0.026817777090602454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,2048,0.018760888112915885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,1536,0.014486221803559197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,8192,0.3597706688774957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,256,0.0032275555034478507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,51200,5.992801666259766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,16384,0.032808889945348106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,768,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,32,0.005321777943107817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,16384,0.07070755296283297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,32,0.01443822185198466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,6144,3.7499237060546875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,128,0.003556444413132138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,64,0.008729777402347988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,4096,0.4051031006707086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,7168,0.6376301977369521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,10240,0.06207999918195936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,51200,0.19535556104448107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,6144,0.013230221966902414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,8192,0.15092889467875162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,16384,8.59739515516493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,128,0.010477333433098264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,16384,0.02346399923165639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,512,0.003914666672547658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,768,0.005551110953092575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,2560,0.1423999998304579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,3072,0.09069422218534683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,1536,0.09316355652279323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,51200,0.01402933398882548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,3072,0.05506222115622627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,8192,2.358740488688151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,768,0.010982222027248807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,8192,0.027466666367318895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,128,0.007722666694058313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,3584,0.5785297817654079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,2048,0.015573332707087198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,65536,0.09848533074061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,65536,4.803005218505859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,3072,0.0324248903327518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,2560,0.009767110976907942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,1024,0.036601778533723615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,16384,0.025154666768179998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,2560,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,7168,0.014697778556081982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,65536,1.8456594679090712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,5120,0.047167999876870044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,2560,0.011912888950771756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,768,0.03489244315359328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,256,0.0047404443224271136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,6144,0.03800355394681295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,4096,0.019174221489164565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,32,0.07065955797831218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,256,0.011727111207114326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,32,0.023660444551044043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,10240,0.0314684444003635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,4096,0.02132622235351139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,4096,0.12727644708421496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,1024,0.07790578073925443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,3072,0.0075493330756823225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,2048,0.059642665916019015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,128,0.005173333403136996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,3072,0.0185208883550432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,128,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,3584,0.009700444837411245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,8192,0.03067644437154134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,65536,0.030991110536787245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,1024,0.008759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,8192,0.15959378083546957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,4096,0.08376444710625543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,10240,0.04224444429079691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,256,0.008063111040327284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,5120,0.09982133573955959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,5120,0.2472648885515001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,1536,0.006737777756320105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,8192,0.09496000077989365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,8192,0.05987199809816149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,64,0.0075804442167282104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,7168,0.30865687794155544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,64,0.005386666705211003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,512,0.007284444239404466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,768,0.010168888502650792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,768,0.007333333293596904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,768,0.0594408909479777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,1536,0.04357599880960253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,128,0.005566222386227713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,5120,0.006913777854707506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,5120,0.02557955516709222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,8192,0.024503111839294434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,8192,0.01331733332739936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,8192,0.2738115522596571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,65536,0.14100978109571669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,512,0.005566222386227713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,8192,0.09303288989596897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,5120,0.015411555767059326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,768,0.04797955685191684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,12288,0.015487111277050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,1536,0.014011555247836642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,64,0.0027075554761621687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,10240,0.05096800128618876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,8192,0.025787555509143408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,5120,0.0890702207883199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,6144,0.017135999268955655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,1024,0.0131217779384719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,16384,1.9496746063232422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,5120,0.06482666730880737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,3072,0.013335111240545908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,10240,0.4059786796569824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,768,0.048764444059795804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,7168,0.3789991007910834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,32,0.00500177757607566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,12288,0.5044533411661783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,1024,0.008792888787057664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,512,0.005205333232879639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,3072,0.021734222769737244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,256,0.009834667046864828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,3584,0.08790489037831624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,768,0.008736888567606607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,768,0.006272000157170826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,65536,0.20582311683230928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,512,0.006695999867386288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,2048,0.005340444544951121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,8192,0.27280089590284556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,10240,0.06511377625995211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,51200,2.5252034929063583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,16384,0.05861600240071615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,128,0.0032266666077905228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,2560,0.021339555581410725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,1536,0.43959111637539333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,16384,0.18733066982693145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,6144,0.3473813268873427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,6144,0.05273422267701891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,2048,0.005597333527273602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,51200,0.03152889013290405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,2560,0.008395555946562026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,512,0.04202311237653097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,3072,0.008003555238246918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,1536,0.008092444803979661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,7168,0.011425777441925473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,512,0.007052444749408298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,1536,0.004451555510361989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,8192,0.02243822150760227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,3072,0.014183110660976835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,768,0.006413333531883027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,256,0.003024000053604444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,1024,0.003975999852021535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,1024,0.007349333001507654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,3584,2.1538658142089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,4096,0.2809608777364095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,7168,0.031016889545652602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,12288,1.400393803914388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,3584,0.01182666669289271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,1536,0.014583110809326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,768,0.010224888722101847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,10240,0.04313244422276815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,1024,0.037618666887283325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,8192,0.19759378168318006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,1536,0.07847377989027235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,512,0.015830222103330824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,12288,0.01444355481200748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,3584,0.7518924607170953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,3584,0.02488888965712653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,3072,0.0155404441886478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,6144,0.3216337892744276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,6144,0.29962754249572754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,2048,0.011988444460762871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,6144,0.22909333970811632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,3584,0.07913244432873197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,12288,0.10052977667914496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,3072,0.017280888226297166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,256,0.004143110993835661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,128,0.016498666670587327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,2560,0.09707378016577826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,768,0.03824533356560601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,5120,0.006232888748248418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,10240,0.07455378108554415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,768,0.005446222093370225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,51200,1.2603555255466037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,7168,0.3191884358723958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,7168,0.014073777529928418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,256,0.029931555191675823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,512,0.0037991110649373797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,256,0.00508355556262864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,65536,10.561417473687067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,128,0.0033573332346147964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,10240,0.025046222739749487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,65536,0.2956373426649305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,6144,0.28450489044189453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,768,0.005640000104904175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,128,0.007278222176763747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,7168,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,768,0.006093333164850871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,3072,0.043918222188949585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,2560,0.009519110951158736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,10240,0.4801955752902561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,2560,0.014496000276671516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,3584,0.008935111264387766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,16384,0.069669332769182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,1536,0.010254222485754225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,5120,0.02234399980968899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,128,0.01901155544651879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,768,0.033019555939568415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,256,0.028969777954949275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,4096,0.039467556609047785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,128,0.15982488791147867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,3072,0.008567999634477828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,4096,0.011221333510345884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,3072,0.011395555403497485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,7168,0.20021066400739881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,512,0.009083555804358589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,16384,0.2613440089755588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,16384,0.06463644239637586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,65536,0.5531333287556967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,64,0.015104000767072042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,8192,0.2158444457583957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,256,0.005469333380460739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,7168,0.08642133076985677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,768,0.0058542220956749385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,12288,0.15882044368320042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,16384,0.02808622188038296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,128,0.007891555627187094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,128,0.002875555513633622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,51200,3.1722763909233938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,6144,5.6816223992241754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,8192,0.43771200709872776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,1024,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,2048,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,512,0.016139555308553908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,65536,0.43387643496195477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,8192,0.04896266592873467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,32,0.008228444390826756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,1536,1.1534542507595487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,3072,3.997541215684679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,1536,1.2854319678412545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,10240,0.08722577492396037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,64,0.004208000169859992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,8192,0.01570222278436025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,16384,0.09103822045856053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,3584,0.03033422099219428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,65536,0.08656977944903904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,2048,0.008046222229798635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,65536,0.2942613230811225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,51200,0.11615910794999863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,10240,0.402312888039483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,65536,0.22243287828233507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,256,0.00462755560874939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,5120,0.012474666866991254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,3072,0.5818053351508247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,4096,0.01314933349688848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,51200,0.138918227619595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,3072,0.011509333219793109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,7168,0.230292452706231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,768,0.012350222302807702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,768,0.061321775118509926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,51200,1.3340631061130102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,12288,0.09554755687713623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,4096,0.06683644321229723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,2048,0.29859201113382977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,32,0.004567999806669024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,51200,0.23105777634514701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,4096,0.041195554865731135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,768,0.06749244530995686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,512,0.005414222263627582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,51200,10.249810112847223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,3584,0.06368533107969496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,2048,0.005306666509972678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,4096,0.13743555545806885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,2048,0.0074577778577804565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,2560,0.007978666987684038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,8192,0.36192621125115293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,1024,3.557086096869575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,768,0.0495013329717848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,12288,0.3655120001898871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,10240,0.3018622133466932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,10240,0.21689511670006645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,4096,0.11488444275326198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,4096,0.19357777966393364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,7168,0.04540355669127571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,65536,0.05955644448598226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,5120,0.009279111193286048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,1536,0.07484977775149874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,512,0.01093777765830358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,7168,0.057484441333346896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,12288,0.04577777783075968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,128,0.005199999858935674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,10240,0.016672000288963318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,2048,0.027123555541038513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,64,0.0029066666546795103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,51200,0.17355999681684706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,12288,0.15194043848249647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,3584,0.04942044284608629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,16384,0.18022667037116158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,3584,0.011757333245542316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,3584,0.1488728920618693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,5120,0.015004444453451367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,512,0.014104889498816596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,7168,0.040779554181628756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,65536,0.20447288619147408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,1024,0.06069066789415148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,2048,0.3334204355875651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,6144,0.05769689215554131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,3072,0.013079110946920184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,6144,0.009962666365835402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,51200,0.07855644490983751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,3072,0.03435466686884562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,2560,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,1536,0.004671999977694618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,2048,0.009840000006887648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,2560,0.013350221845838757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,32,0.006066666709052191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,256,0.00720711135201984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,6144,2.1394293043348527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,3072,0.018420444594489205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,1536,0.006682666639486949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,10240,0.17798754904005262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,65536,0.36001155111524796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,65536,0.1920320060518053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,512,0.020571556356218126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,32,0.006036444256703059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,65536,0.1381937795215183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,2560,0.03526666760444641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,4096,0.011078221930397881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,51200,10.897668626573351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,10240,0.9613093270195855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,16384,0.4279279973771837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,6144,0.018009778526094224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,51200,5.336914486355251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,2048,0.009116444322797988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,10240,0.6528524292839898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,512,0.004908444566859139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,10240,0.11134221818712021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,128,0.0028648889727062653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,3072,0.08618933624691433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,768,0.0038328886859946777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,5120,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,12288,0.3405662112765842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,51200,0.11709600024753147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,768,0.019478221734364826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,3072,0.3327377902136909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,512,0.13981066809760198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,7168,0.1841599941253662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,12288,0.031679110394583807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,12288,0.12482399410671657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,256,0.20665421750810412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,2560,0.10928089088863796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,16384,0.055404444535573326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,1024,0.9862346649169922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,128,0.004368888835112254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,2048,0.07757511403825548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,2560,0.010307555397351583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,2048,0.007045333584149678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,3584,0.04773599902788798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,1024,0.006055999961164262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,3584,0.1517662207285563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,51200,0.02327288852797614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,4096,0.1081048912472195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,768,0.02316088808907403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,512,0.004911111046870549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,16384,0.03670577870474921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,10240,0.024923556380801733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,7168,0.005967999911970562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,12288,0.15613778432210287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,1024,0.035953776703940496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,512,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,8192,0.2780311107635498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,10240,0.25117866198221844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,65536,0.46006133821275497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,8192,0.02019999921321869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,6144,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,10240,0.01847466660870446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,65536,0.0364951127105289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,4096,0.49232620663113064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,128,0.004375111311674118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,128,0.00499288903342353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,3584,0.0069324444565508105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,3072,0.01387466655837165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,32,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,4096,0.013387555877367655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,8192,0.04465155469046699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,2048,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,3584,0.023639111055268183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,2048,0.008247111406591203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,5120,1.2551715638902452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,5120,0.01863466699918111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,16384,0.24645511309305826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,64,0.011706666813956367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,51200,0.31871377097235787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,51200,0.02697244452105628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,7168,0.041645334826575384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,51200,0.0541928874121772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,768,0.003847111016511917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,51200,0.6783750851949056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,256,0.003343111111058129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,16384,0.022043555974960327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,1024,0.00981244444847107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,4096,0.10696799887551202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,32,0.0027306667632526825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,128,0.009975999593734741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,16384,0.16448711024390325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,6144,0.043232887983322144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,512,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,3072,0.0120666664507654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,768,0.003628444340493944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,65536,1.0600426991780598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,7168,0.1250764396455553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,512,0.05821777714623345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,4096,0.03872177667087979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,7168,0.020471110939979553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,65536,0.10609689023759629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,8192,0.006797333144479328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,2048,0.012643555800120035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,1536,0.01056177748574151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,4096,1.6400089263916016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,10240,0.07198755608664618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,768,0.004121777911980947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,12288,0.10512533452775742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,512,0.007118222614129384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,512,0.005716444303592046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,3072,0.025934222671720717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,12288,0.17007377412584093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,512,0.006007110906971826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,1536,0.13655555248260498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,768,0.3230302333831787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,2560,0.08331288894017537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,64,0.00572533326016532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,256,0.020807110601001315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,12288,0.030785779158274334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,2560,0.010781333678298526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,768,0.003809777812825309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,3584,0.2014168898264567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,6144,0.02461333407296075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,2048,0.23608178562588164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,16384,2.011076397365994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,5120,0.05540533198250664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,32,0.010127999716334874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,51200,0.10492622190051609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,51200,0.014592889282438489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,7168,1.3102195527818468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,4096,0.021273778544531927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,128,0.004600888739029567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,10240,0.649671130710178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,4096,0.01610133383009169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,4096,0.18445777893066406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,512,0.013248888982666863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,1024,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,2560,0.042767110798094005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,768,0.04557333389918009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,10240,0.6217715475294325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,3584,0.00962933318482505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,3584,0.00908622228437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,2048,0.022278222772810195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,51200,0.11147822274102105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,1024,0.023245332969559565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,10240,0.34119378195868594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,5120,0.028047998746236164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,768,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,1024,0.010264889233642155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,7168,0.35155465867784286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,65536,0.5910559760199653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,2560,0.044963555203543774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,12288,0.010660444696744284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,256,0.027320888307359483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,12288,3.4491297403971353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,51200,0.5917964511447483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,7168,0.2336026562584771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,256,0.00433955548538102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,1536,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,768,0.025965332984924316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,65536,36.57530042860243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,12288,0.024518221616744995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,8192,0.01644622286160787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,5120,0.17493689060211182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,12288,0.06203022268083361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,64,0.04818577898873223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,768,0.010475555227862464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,32,0.059664891825781934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,3072,0.00977422214216656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,10240,0.03882844580544366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,512,0.004639111045334074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,65536,0.059511111842261426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,3072,2.1691955990261502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,8192,0.022341334157519873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,10240,0.03962666789690653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,768,0.010954666468832227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,128,0.0031831111345026228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,512,0.015554666519165039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,8192,0.10371733374065822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,7168,0.020990222692489624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,256,0.015797333584891427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,8192,0.1484231154123942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,128,0.006631999793979857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,7168,0.14848355452219644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,128,0.004416888786686791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,6144,0.02253244486120012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,4096,0.023159111539522808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,51200,0.7490640216403537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,16384,0.02423200011253357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,3584,0.01184088902340995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,51200,0.05564000209172567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,10240,0.17310755782657197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,65536,0.3702204492357042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,6144,0.06649066342247857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,1536,0.006224000205596288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,10240,0.6302995681762695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,10240,0.04467022087838915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,8192,0.17876799901326498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,32,0.16711555586920845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,12288,0.06781422429614596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,51200,1.4974418216281469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,16384,0.02546577817863888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,2560,0.046662221352259316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,8192,0.22850754525926378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,7168,0.23781243960062662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,16384,0.050245334704717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,3584,0.18455289469824898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,8192,0.12477689319186741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,12288,0.09512266847822402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,8192,0.01959377858373854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,3584,0.11685866779751247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,10240,0.2155973381466336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,1536,0.05873066849178738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,10240,0.016089777151743572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,256,0.008457777400811514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,768,0.02164266672399309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,64,0.0026515554636716843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,1024,0.012260444462299347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,512,0.0053680001033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,64,0.002581333327624533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,4096,0.03144355614980062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,4096,0.03719733489884271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,10240,0.06928444570965238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,512,0.008010666403505538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,768,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,64,0.007160888777838812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,8192,0.11339022053612603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,64,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,3072,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,7168,0.02741777731312646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,1024,0.007791111038790808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,1536,0.0225600004196167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,51200,0.08598133590486313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,51200,0.7595226499769422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,2048,0.060989333523644336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,32,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,512,0.003256888853179084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,3584,0.015433778365453085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,12288,0.032999111546410456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,64,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,3584,0.1285982264412774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,1024,0.4581555790371365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,16384,0.06957244210773043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,128,0.11340088314480251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,1536,0.03164711263444688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,10240,0.01976799964904785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,3072,0.05197155475616455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,12288,0.008231999973456064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,5120,0.6626008881462945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,64,0.0034497777620951333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,3584,0.018014222383499146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,512,0.024088889360427856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,32,0.014516444669829475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,3072,0.008352888955010308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,4096,0.013599110974205865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,1024,0.10287200080023871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,256,0.00721777809990777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,4096,0.23462311426798502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,65536,0.2103137837515937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,2048,0.005504000104135937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,3072,0.4454639752705892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,65536,1.7925137413872614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,8192,0.23252799775865343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,7168,0.012625777886973487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,5120,0.016699555847379897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,128,0.017835555805100333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,8192,0.12405955791473389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,768,0.1191377772225274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,10240,0.21105866962009004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,12288,0.07090133428573608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,512,0.004392888810899523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,10240,0.013431111143694984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,12288,0.01830933325820499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,65536,0.21930844253963896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,3584,0.059989333152770996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,64,0.0035786665976047516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,65536,0.04168533285458883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,51200,0.007640000018808577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,32,0.002528000002106031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,32,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,4096,0.014087110757827759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,7168,0.0262106673585044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,7168,0.03218844532966614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,16384,0.06551021999782987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,768,0.07613066832224528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,6144,0.0158595558669832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,4096,0.026338666677474976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,5120,0.372292439142863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,128,0.002793777734041214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,65536,0.1683795584572686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,16384,0.10028444396124946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,6144,0.6056497891743978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,768,0.025452444950739544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,10240,0.048584000931845776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,16384,40.26803249782986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,6144,0.05743910868962606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,65536,0.023030221462249756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,128,0.010125333236323463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,7168,0.12657066186269125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,256,0.11296978261735703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,65536,0.44853957494099933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,1024,0.004077333129114575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,7168,0.012145777543385824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,1536,0.004810666872395409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,1024,0.009855111440022787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,4096,0.06759555472267999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,7168,0.011052444577217102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,7168,0.37160444259643555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,3584,0.013721778160995908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,512,0.005514666852023866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,8192,0.015896000795894198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,51200,2.886590321858724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,8192,0.1265182230207655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,10240,0.14118844932980007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,51200,0.9128888448079427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,10240,0.04720800121625265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,2560,0.1804577774471707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,16384,0.018036444981892902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,7168,0.023413333627912734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,2048,0.006224000205596288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,12288,0.2891937891642253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,10240,0.0490497781170739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,768,0.48275110456678605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,12288,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,16384,0.02713777787155575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,128,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,64,0.005320000151793162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,1024,0.015219555960761176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,12288,0.016354666815863717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,256,0.019547555181715224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,3072,0.07336888710657756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,65536,0.9595573213365344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,3584,0.013182222015327878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,3072,0.015555555621782938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,65536,0.9426666895548502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,2560,0.40751200252109104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,4096,0.014000000225173103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,12288,0.5228293206956651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,2048,0.0188746667570538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,7168,0.023473777704768713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,2048,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,256,0.005422222117582957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,256,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,10240,0.04813244607713488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,6144,0.3218657705518934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,256,0.004952888935804367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,6144,0.05510310994254219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,3072,0.19242221779293486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,7168,0.18457689550187853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,3072,1.0508284038967557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,128,0.0307004451751709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,1536,0.007838221887747446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,12288,0.9167502721150717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,8192,0.01629866659641266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,8192,0.007891555627187094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,2048,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,7168,0.4783422152201335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,1024,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,32,0.007867555651399825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,3072,0.009583111438486312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,3072,0.010832889212502373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,2560,0.2227706644270155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,12288,0.05517333414819506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,4096,0.06688888867696126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,32,0.035248888863457575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,4096,0.008951110972298516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,3072,0.01402933398882548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,4096,0.18107288413577607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,5120,4.115489959716797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,10240,0.020752888586786058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,65536,5.133992089165582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,2048,0.010403555300500657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,128,0.005708444449636672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,2560,0.09442577759424846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,4096,0.017640888690948486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,12288,0.030188444587919448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,3072,0.022327999273935955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,512,0.036730666955312095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,3072,0.05650222301483154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,16384,0.016488000750541687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,2048,0.015327110886573792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,7168,1.0033146540323894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,4096,0.06487288739946154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,32,0.005054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,5120,0.09732533163494533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,32,0.015554666519165039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,768,0.006162666612201267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,4096,0.011526222030321756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,4096,0.019727999965349834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,128,0.005972444597217772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,4096,1.0822622511121962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,6144,0.01631733278433482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,512,0.03687999977005853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,512,0.003597333199448056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,8192,0.9472106297810873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,6144,0.21458844343821207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,65536,0.8380631340874566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,2048,0.07279466920428805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,51200,0.17430933316548666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,5120,0.012785777449607849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,2560,0.007970666719807519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,3072,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,6144,0.02758399976624383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,10240,0.015625778171751235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,2048,0.02607466611597273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,3072,0.038584887981414795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,4096,0.09533955653508504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,5120,0.011432888607184092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,6144,0.016116445263226826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,10240,0.2601982222663032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,5120,0.01440533333354526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,32,0.05948533614476522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,512,0.01401600076092614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,65536,0.49507464302910703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,512,0.004835555536879433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,5120,0.025038222471872967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,2048,0.019890666007995605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,3072,0.07383200195100573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,64,0.011262222296661802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,128,0.007877333296669854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,5120,0.018439110782411363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,1024,0.0057884446448749965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,10240,0.0812551114294264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,2560,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,10240,0.021611554755104914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,8192,0.16165067089928523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,3584,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,2560,0.018616888258192275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,128,0.002817777709828483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,512,0.003809777812825309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,768,0.022786666949590046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,7168,0.1253733370039198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,3584,0.008069333102968004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,6144,0.025231111380789015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,1024,0.008397333323955536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,32,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,1536,0.034894221358829074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,6144,0.07306844658321805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,5120,2.336259629991319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,4096,0.3330106735229492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,7168,0.3788435459136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,1536,0.006959111326270633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,16384,0.6456640031602647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,8192,0.014767999450365702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,7168,0.021922666165563796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,512,0.004101333518822988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,3072,0.11145955986446804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,768,0.018974221414989896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,2560,0.19205333126915825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,1024,0.04788800080617269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,4096,0.009298666483826106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,5120,0.011362666885058085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,128,0.01796533332930671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,7168,0.5476595560709635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,51200,0.22951555252075195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,512,0.005345777918895085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,12288,0.10798933108647664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,256,0.04028444488843282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,5120,0.01051644484202067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,3072,0.012229333321253458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,12288,0.5896257824367946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,256,0.00656088896923595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,64,0.11978933546278213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,256,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,512,0.00489688871635331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,1024,0.05362755391332838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,65536,0.9170142279730903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,6144,0.029320889049106177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,1536,0.022030222747060988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,1536,0.007221333682537079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,5120,0.016864000095261466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,1024,0.0385484430525038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,5120,0.010384000009960597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,256,0.01368444412946701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,6144,0.05503377649519178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,65536,0.33798578050401473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,10240,0.039111998346116804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,65536,0.1491031116909451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,10240,0.7031546698676215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,2048,0.053307556443744235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,3072,0.027658666173617046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,128,0.004171555654870139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,2560,0.01037333326207267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,1536,0.0321777794096205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,6144,0.03772177630000644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,32,0.002752888947725296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,1024,0.005842666659090254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,51200,0.05487822161780464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,2048,0.00926399976015091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,8192,0.020018666982650757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,2560,0.015358222855461968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,7168,0.23663645320468477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,8192,0.013570666313171387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,6144,0.2673848999871148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,6144,0.010258666343159145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,6144,0.010888888604111142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,32,0.05735466877619425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,3072,0.10749866565068562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,2560,0.07610844241248237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,128,0.004456888884305954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,7168,0.962100452846951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,4096,0.027256889475716486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,4096,0.011521778172916837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,16384,0.9083502027723523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,10240,0.4165368874867757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,2048,0.06551466385523479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,4096,0.05513866742451986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,256,0.006176888942718506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,10240,0.18966489368014863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,51200,1.3350053363376195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,16384,0.29573512077331543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,8192,0.014320888453059726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,2560,0.016935111747847665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,2048,0.007094221810499827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,51200,0.30174133512708873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,5120,0.022414222359657288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,16384,0.029126223590638902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,10240,0.006493333313200209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,5120,0.2100515498055352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,5120,0.009576000273227692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,16384,0.14051022794511583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,3584,0.06352444489796956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,51200,0.10405777560340033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,4096,0.006340444501903322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,10240,0.06334311432308622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,3072,0.058471997578938804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,12288,0.7714311281840006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,51200,6.655970679389106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,2048,0.08751733435524835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,768,0.007826666865083907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,10240,0.023930667175187006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,2048,0.012709333664841123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,12288,1.8312933180067275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,10240,0.07358133130603366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,1536,0.1145813332663642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,1536,0.03393155667516921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,7168,0.012576888832781049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,10240,0.02860800094074673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,3584,0.008827555510732863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,2560,0.0069102222720781965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,2048,0.01647733317481147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,16384,0.021548445026079815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,16384,0.12742577658759222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,2048,0.012244444754388599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,32,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,768,0.017664000391960144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,51200,0.0543084442615509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,1024,0.3804337713453505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,8192,1.3409173753526475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,256,0.05517066518465678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,768,0.006096000058783426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,1536,0.014946666028764514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,512,0.004971555537647671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,51200,0.3062266773647732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,3584,0.18348088529374865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,6144,0.023964444796244305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,4096,0.011081777513027191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,128,0.0027751111321979095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,10240,0.4147831069098578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,5120,0.03166844447453817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,8192,1.1863750881618924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,512,0.02892000145382351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,12288,0.23379288779364693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,6144,0.024000888069470722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,12288,0.20515733295016816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,32,0.007090667055712805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,512,0.005797333187527127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,1024,0.005088888936572605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,1024,0.02809422214825948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,6144,0.013170666992664337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,12288,0.20834844642215303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,6144,0.013999111122555204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,10240,0.01944266590807173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,10240,0.01940711173746321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,768,0.005434666656785541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,8192,0.01148977792925305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,5120,0.009312888814343346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,128,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,7168,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,1536,0.008395555946562026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,65536,2.771663029988607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,128,0.003834666477309333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,512,0.004488888714048597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,16384,0.49809779061211484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,5120,0.042508443196614586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,512,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,2048,0.011490667031870948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,768,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,4096,0.007333333293596904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,7168,0.040785776244269475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,512,0.01773155563407474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,32,0.004139555411206352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,51200,0.0913955569267273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,128,0.019647111495335896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,2560,0.00812977800766627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,2560,0.008437333007653555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,3584,0.011603555745548673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,4096,0.02888710962401496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,2560,0.013722666435771518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,10240,0.6507724655999078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,12288,1.3704862594604492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,7168,0.06335644589530097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,3072,0.03544622328546312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,64,0.003302222324742211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,32,0.009855111440022787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,256,0.006346666564544042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,5120,0.013269333375824822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,3584,1.0425466961330838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,51200,4.532388475206163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,4096,0.03651555710368686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,2560,0.04900622367858887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,768,0.04407555527157254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,2560,0.03768710957633124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,512,0.01682222220632765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,12288,0.3538293308681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,1024,0.19345688819885254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,128,0.00906488878859414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,2048,0.16086310810512966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,4096,0.7666222254435221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,4096,0.05008888906902737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,6144,0.12204533153110081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,7168,0.0159582214223014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,8192,0.017124444246292114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,128,0.08448177576065063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,51200,0.03410844339264764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,8192,0.03903199897872077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,51200,1.9823307461208766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,3072,0.008926221893893348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,128,0.006919999917348226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,12288,0.019111999207072787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,16384,0.039657778210110135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,768,0.09580800268385146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,2560,0.03868622250027127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,2560,0.016007999579111736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,12288,0.012666666673289405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,4096,0.007682667010360294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,4096,0.008197333249780867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,5120,0.043954667117860585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,128,0.005984888722499211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,32,0.003067555526892344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,256,0.004416000097990036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,7168,0.01294133315483729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,256,0.011615110768212212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,16384,0.057856890890333385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,2560,0.12773599889543322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,768,0.023260444402694702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,128,0.01537333263291253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,10240,0.0370124446021186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,768,0.006603555546866522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,5120,0.019733332925372653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,2560,0.007092444433106317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,51200,0.0527671111954583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,256,0.003096888876623578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,6144,0.04112889038191901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,10240,0.05245866709285312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,6144,2.123328102959527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,128,0.01091733326514562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,3072,0.020432000358899433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,128,0.01404622197151184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,5120,0.02016888890001509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,65536,0.056686222553253174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,16384,0.008476444416575963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,128,0.0035173334181308746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,256,0.011952000359694162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,1536,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,1536,0.007101332975758447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,4096,0.1765902174843682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,5120,0.04126222266091241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,3072,0.009972444838947719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,128,0.011320888996124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,5120,0.009941333697901832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,65536,0.0169813334941864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,12288,0.043466667334238686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,1536,0.01309511148267322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,256,0.011750222080283694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,128,0.08272711435953777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,2048,0.019505777292781405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,8192,0.08583555618921916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,4096,0.015229332778188916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,768,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,2048,0.37723467085096574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,16384,0.6738799942864312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,65536,0.19130310747358534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,32,0.005683555371231503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,4096,0.028631998433007136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,512,0.003559999995761447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,6144,0.05610311031341553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,2560,0.07447910971111722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,65536,0.8659608629014758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,2560,0.010652444428867765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,32,0.008287111090289222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,5120,0.10023911131752862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,3584,0.01790844400723775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,16384,0.00647288892004225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,51200,0.11044177744123672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,1024,0.02331377731429206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,256,0.015479111009173922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,51200,0.42370221349928117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,7168,0.0307279990779029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,2560,0.006019555446174409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,7168,0.011319999893506369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,8192,0.007139555282062954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,65536,0.10684266355302598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,3584,0.07533333036634657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,768,0.0069671107663048645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,1536,0.004621333132187526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,8192,0.01498222185505761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,12288,0.30447377098931205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,3584,0.06502666738298205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,1536,0.0772542224989997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,5120,0.05808089176813761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,4096,0.0070568886068132185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,2560,0.5404177771674262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,1024,0.010262221925788457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,51200,0.28172267807854545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,2048,0.011352000137170156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,2560,0.061230222384134926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,32,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,64,0.0025635556214385563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,1536,0.00675999994079272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,4096,0.01551466683546702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,3584,0.14446132712894016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,256,0.0874168872833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,8192,0.006296000132958095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,32,0.015236443943447538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,256,0.02834577692879571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,512,0.19940000110202363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,10240,0.01538222200340695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,10240,0.305768887201945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,768,0.005992888990375731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,64,0.0033617777129014335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,51200,0.0605857769648234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,512,0.23844443427191842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,128,0.0029528888149393927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,4096,0.0088755554623074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,4096,0.007923555870850882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,1024,0.04494666722085741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,32,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,3584,0.01013777818944719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,2048,0.010429333481523724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,2560,0.4523608949449327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,6144,0.060962663756476514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,5120,0.07757421996858385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,64,0.002554666664865282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,2048,0.008579555484983657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,65536,0.053595556153191455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,3072,0.05062311225467258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,32,0.002733333243264092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,12288,0.04926133486959669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,3072,0.08884177605311076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,3584,0.008004444340864817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,64,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,64,0.01770666738351186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,3072,0.010139555566840703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,2048,0.006607999818192587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,1024,0.01965777741538154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,2560,0.05487377776039971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,51200,1.1223235660129123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,6144,0.01534844438234965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,12288,0.006723555425802867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,3072,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,768,0.01726222203837501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,4096,0.009840889109505547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,3072,0.1373200019200643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,2048,0.01147555559873581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,65536,1.1423217985365126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,5120,0.010953777366214328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,3584,0.06256444586647882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,3584,0.013376889129479727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,65536,5.211661444769965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,65536,1.2221626705593533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,4096,0.010544000400437249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,128,0.003440888805521859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,768,0.007455999652544658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,16384,0.02966044346491496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,8192,0.02755733331044515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,6144,0.014330666926172046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,8192,0.02145511077509986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,65536,0.20446399847666422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,5120,0.029900444878472224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,6144,0.19776977433098686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,7168,0.32757067680358887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,128,0.01770666738351186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,10240,0.1485662195417616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,32,0.004384888956944148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,8192,2.262889862060547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,12288,0.16489156087239584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,10240,3.056807200113932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,2560,0.00924355536699295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,4096,0.039451556073294744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,6144,0.01864266726705763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,1024,0.016725333200560678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,7168,0.026485333840052288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,1536,0.008405333591832055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,2048,0.025987555583318073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,3584,0.020950223008791607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,4096,0.04450133442878723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,4096,0.018226666582955253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,5120,0.009344889058007134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,32,0.009710222482681274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,10240,0.2674364513821072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,7168,0.04383644461631775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,7168,0.05658577548133003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,10240,0.3105439874860975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,6144,0.4482391145494249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,512,0.007039111521508958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,3072,0.014959999256663851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,4096,0.08277510934405856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,1536,0.013630222115251752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,65536,1.8284621768527562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,16384,0.2947946654425727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,128,0.004310222135649787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,3584,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,768,0.638296021355523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,2048,0.05400266581111484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,1024,0.006367110957702001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,8192,0.06673866510391235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,16384,0.04937066634496053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,512,0.060030221939086914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,7168,0.016318221886952717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,4096,0.01937955617904663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,5120,0.13091466161939833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,3072,0.014352889524565803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,32,0.011696000066068439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,3072,0.07784710990058051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,8192,0.14012000295850965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,2560,0.01088711122671763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,65536,2.293307622273763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,5120,0.06326933039559259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,2560,0.008389333056079017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,3584,0.012342222034931183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,65536,0.07830488681793213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,51200,16.731970893012154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,7168,0.01519555515713162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,3584,0.17484622531467012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,2048,0.47308980094061953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,768,0.015145777000321282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,2560,0.06286222404903836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,2560,0.009129777550697327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,2560,0.014095111025704278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,768,0.008801778157552084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,8192,0.20622489187452528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,1536,0.00675644435816341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,65536,0.08305955595440336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,65536,0.47707197401258683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,64,0.18168621593051484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,10240,0.19763288233015272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,65536,1.2825795279608834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,5120,0.09485866626103719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,128,0.006575999988449945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,16384,0.15142667293548584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,10240,0.015619556109110514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,7168,0.04544088906712002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,12288,0.20880977312723795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,1024,0.006614222294754452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,3072,0.012050666742854647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,51200,0.3322977754804823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,3072,0.050251556767357715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,128,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,4096,0.011775111158688864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,3584,0.10050400098164876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,512,0.02295911146534814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,7168,0.014085332552591959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,51200,26.473982069227432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,3072,0.01254488858911726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,8192,0.03698044353061252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,3072,0.20635199546813965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,8192,0.01811022228664822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,2048,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,5120,0.007745777567227681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,7168,0.08163199822107951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,6144,0.1589555607901679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,768,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,10240,0.09396977557076348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,65536,0.9887768427530924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,768,0.004328888737493091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,64,0.006171555568774541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,4096,0.01204533295498954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,8192,0.02015911042690277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,16384,4.504981146918403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,768,0.013031110995345645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,7168,0.07448977894253202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,64,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,2048,0.00666844430896971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,1536,0.01185155577129788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,256,0.037243554989496865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,8192,0.0717173351181878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,2048,0.0184871107339859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,16384,0.02681422233581543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,10240,0.016696888539526198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,1024,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,5120,0.0850871139102512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,16384,2.104502148098416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,1024,0.007850666840871176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,64,0.008218666745556725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,12288,0.025220443805058796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,16384,0.28450576464335126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,6144,0.01643200053109063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,2560,0.18222578366597494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,64,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,1536,0.004406222038798862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,7168,0.047566221819983594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,51200,0.07282488875918917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,3072,0.01144088887506061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,3584,0.007432888779375289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,64,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,10240,0.16467021571265325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,8192,0.14597511291503906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,1024,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,16384,0.01959377858373854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,32,0.003343111111058129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,6144,0.016750221451123554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,768,0.006239999913507038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,10240,0.2919982274373372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,5120,0.21879377630021837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,3584,0.11568000581529404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,512,0.007602666815121968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,512,0.0046880000995265115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,128,0.0027875554644399216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,51200,0.07031289074156019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,2048,0.00812000036239624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,256,0.005355555564165115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,4096,0.008753777378135258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,10240,0.05126133229997423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,10240,0.04061422083112929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,512,0.028939555088678997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,2048,0.07408977879418267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,4096,0.008191111187140146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,8192,0.21442843808068168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,64,0.002986666642957263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,2560,0.12341245015462239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,256,0.013831111292044321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,1024,0.0046657779150538975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,10240,0.016571554872724745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,2048,0.019609777463806998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,1536,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,8192,0.08457422256469727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,768,0.005176888985766305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,6144,0.013235555754767524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,128,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,1536,0.05228266782230801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,6144,0.32415821817186147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,256,0.007242666350470648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,32,0.01717688971095615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,4096,0.08854133552975124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,768,0.015007111761305066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,5120,0.015013333823945789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,51200,0.27680444717407227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,16384,0.24984976980421278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,65536,1.7398950788709853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,7168,0.02347911066479153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,16384,0.022675555613305833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,65536,0.10928799708684285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,32,0.005761777775155173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,32,0.005092444519201915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,32,0.0025724443710512584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,10240,2.4592906104193792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,1024,0.01826133330663045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,10240,0.024293333292007446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,768,0.07750311162736681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,51200,0.07048355870776706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,128,0.004380444271696938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,512,0.013202666408485837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,32,0.005306666509972678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,6144,0.15972177187601724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,12288,0.032961779170566134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,2560,0.009631111390060848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,1536,0.00572533326016532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,3072,0.02632444434695774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,7168,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,256,0.04610133171081543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,256,0.17334844006432426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,32,0.007119110888904995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,2560,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,8192,0.021344888541433547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,6144,0.027777777777777776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,16384,0.3322969012790256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,1024,0.00958933350112703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,128,0.0033982222278912864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,64,0.0028071111689011254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,3072,0.04601066642337375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,10240,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,2560,0.09391199880176121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,10240,0.04565155506134033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,6144,0.07487644751866658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,128,0.005043555465009478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,16384,0.48492177327473956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,65536,0.0887297789255778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,768,0.10903200176027085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,10240,0.10276089111963908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,10240,0.12211999628278945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,8192,0.06939022408591376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,10240,0.8473502265082465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,10240,1.1023164325290256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,3072,0.6116213268703884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,3584,0.11539999643961589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,1024,0.022651554809676275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,65536,0.7461244265238444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,7168,0.09890133142471313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,6144,0.012615111139085559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,128,0.022557333111763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,256,0.005597333527273602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,64,0.0025635556214385563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,65536,0.5919982062445747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,65536,1.673721737331814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,4096,0.039530667993757464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,3584,0.06842399968041314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,3072,0.022348443667093914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,2560,9.332809448242188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,512,0.011623111036088733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,10240,0.5002391073438857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,5120,0.04829600122239855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,4096,0.07892800039715238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,64,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,4096,0.07565066549513075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,65536,5.377441830105251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,3584,0.09852711359659831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,1024,0.011368888947698804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,12288,0.02186044388347202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,256,0.003506666670242945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,65536,0.08011111285951403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,10240,0.03541244400872125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,16384,2.0354506174723306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,6144,0.051685333251953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,256,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,2560,0.019003555178642273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,7168,0.015235554840829639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,6144,0.2019431061214871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,1536,0.021126222279336717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,4096,0.012687111066447364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,8192,0.12699377536773682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,7168,0.027475555737813313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,65536,0.31843821207682294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,6144,0.015071999695565967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,3584,0.022287999590237934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,65536,1.840311050415039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,128,0.03776889046033224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,512,0.17067555586496988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,6144,0.10668444633483887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,4096,0.012618666721714867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,7168,1.7922479841444228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,32,0.017755554782019723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,5120,0.025820444027582806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,3584,0.430261320537991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,6144,0.26207200686136883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,4096,0.03292622168858846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,10240,0.08367022540834214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,64,0.005347555710209741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,5120,0.25629689958360463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,512,0.008178667061858708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,64,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,6144,0.022524444593323603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,2560,0.08154222038057116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,5120,0.013272888958454132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,6144,0.2535937892066108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,2048,0.015215999550289579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,1536,0.21389688385857475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,4096,0.14989066123962402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,512,0.0058906666106647914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,2048,0.009758222434255812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,8192,0.011915555430783166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,2560,13.645791795518663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,6144,0.017116443978415597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,65536,0.10661955674489339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,6144,0.10495999786588882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,8192,0.029706666866938274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,3072,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,2048,0.0070017774899800616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,2048,0.007259555160999298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,1536,0.007301333049933116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,8192,0.7434728940327963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,32,0.005984888722499211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,10240,0.018368000785509746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,8192,0.023927110764715407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,7168,0.028657777441872492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,65536,1.1930533515082467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,32,0.002969777832428614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,1536,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,51200,0.17332000202602812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,51200,0.17325777477688262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,5120,0.009786666267448002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,768,0.008739555875460306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,4096,0.010435555544164447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,1536,0.015150222513410779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,1024,0.042613334125942655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,6144,0.017271111408869427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,12288,0.3776622348361545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,3072,0.06462399827109443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,32,0.014884443746672736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,2560,0.026547556122144062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,6144,0.13557867209116617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,65536,2.607961866590712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,64,0.00869688888390859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,2048,0.15998933050367567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,5120,0.01371911085314221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,3584,0.027023111780484516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,64,0.00555377784702513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,7168,0.014456889695591398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,12288,0.23684800995720756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,1536,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,6144,0.009140444298585257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,6144,0.006256888724035687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,7168,0.19512532816992867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,256,0.03442400031619602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,128,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,1024,0.005264889034960005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,3584,0.14253600438435873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,1536,0.006776888751321369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,3584,0.026837334036827087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,2560,0.08967466486824884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,65536,1.7079920238918727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,1536,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,1024,0.0637600024541219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,16384,1.8858587476942275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,4096,0.5745013554890951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,128,0.27843909793429905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,32,0.004934222333961063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,1536,0.014917333920796713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,32,0.002736888825893402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,16384,0.024414221445719402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,3584,0.04916266600290934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,8192,0.1703235573238797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,12288,0.0269813338915507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,4096,0.09080177545547485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,2560,0.015104000767072042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,512,0.05527199970351326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,1536,0.06855199734369914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,128,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,8192,0.011761777930789523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,32,0.003959111041492886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,2048,0.0736408895916409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,1536,0.012995555169052549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,256,0.003879111260175705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,256,0.0369031098153856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,512,0.0229102224111557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,7168,0.04386577672428555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,1536,0.00591822216908137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,65536,2.306403477986654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,7168,0.022114665971861944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,1024,0.4104124440087213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,10240,0.05780533287260267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,6144,0.013854222165213691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,16384,0.01793333391348521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,4096,0.059346669250064425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,1536,0.7187466621398926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,8192,0.05964889129002889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,3072,0.2997671233283149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,2048,0.21562488873799643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,1536,0.008808888494968414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,6144,0.46502134535047746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,5120,0.04174400038189358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,768,0.039656887451807656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,64,0.0027582223216692605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,2048,0.08109155628416273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,10240,0.027862221002578735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,32,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,2048,0.011595555477672152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,16384,0.020237333244747586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,128,0.004109333372778363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,32,0.005025777965784073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,512,0.03305866652064853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,6144,0.023735110958417256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,6144,0.03552622265285916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,4096,0.020968889196713764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,65536,0.20124533441331652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,6144,0.0099973330895106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,256,0.006989333364698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,1536,0.015683554940753512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,64,0.15285510487026638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,768,0.006203555398517185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,256,0.005028444445795483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,8192,0.27986666891309947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,10240,0.21919022666083443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,128,0.002895111011134254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,512,0.03752800159984165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,512,0.01590666671593984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,1024,0.028821332587136164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,256,0.031918221049838595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,256,0.04367466767628988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,8192,0.10519733031590779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,10240,0.32639821370442706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,7168,0.05897421969307793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,256,0.023385778069496155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,65536,0.07379200061162312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,12288,0.017121778594122995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,4096,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,3584,0.026191110412279766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,10240,1.1604844199286568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,1024,0.27926932440863717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,256,0.003963555726740095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,65536,4.304681566026476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,3584,0.023992889457278784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,128,0.003412444351447953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,1024,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,65536,0.3860035472446018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,12288,0.028625776370366413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,3584,0.011808888779746162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,768,0.005884444548024072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,16384,0.3047093285454644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,3072,0.08710577752855088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,5120,0.1353315512339274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,51200,0.11359288957383896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,256,0.0038284444146686127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,64,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,65536,4.192542182074653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,5120,0.02597066594494714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,16384,0.03804622093836466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,3584,0.00982844415638182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,2560,0.00866844422287411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,16384,0.13669333193037245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,7168,0.008282666405042013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,5120,0.016530666086408827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,10240,0.02828888760672675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,32,0.0028106667515304354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,32,0.0028239999794297745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,3072,0.015075556106037565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,4096,0.03344977895418803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,128,0.01204533295498954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,4096,0.019627556204795837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,7168,0.011030221978823343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,1024,0.003991999973853429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,1536,0.06282044119305082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,5120,0.01863466699918111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,256,0.0032826666202810076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,768,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,51200,0.13387110498216417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,65536,0.04892355534765455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,8192,0.05838577614890205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,128,0.013219555219014486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,128,0.0042035554846127825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,1536,0.025607110725508794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,16384,0.324743111928304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,2560,0.02070488863521152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,5120,0.5878204239739312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,51200,0.009123555488056606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,2048,0.04178399840990702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,8192,0.018368888232443068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,51200,0.09204977750778198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,2048,0.019127999742825825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,1536,0.00628977765639623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,3072,0.15738577312893337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,1024,0.0051857779423395796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,51200,0.10345866945054795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,1536,0.007032000356250339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,1536,0.013154666456911298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,1536,0.009927999642160203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,2560,0.028949333561791316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,128,0.008375110725561777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,2048,0.07749244239595202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,1536,0.01479111115137736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,12288,0.006512889017661412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,2048,0.5195866690741645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,65536,0.03686133358213637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,10240,0.043972445858849414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,128,0.005208888815508948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,2048,0.006208000083764394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,5120,0.032968888680140175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,768,0.008784888519181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,64,0.0028737777223189673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,1024,0.007355555891990662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,7168,0.013585777746306526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,3584,0.008246222303973304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,7168,0.06495644648869832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,128,0.00309599998096625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,3584,0.03367822368939718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,256,0.005029333134492238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,64,0.004843555390834808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,6144,0.09937155246734619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,64,0.011737777955002256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,16384,0.02286311156219906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,256,0.008971555365456475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,768,0.01906755566596985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,1536,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,2560,0.006055999961164262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,512,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,64,0.020204444726308186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,16384,0.021418665846188862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,6144,0.20959556102752686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,3584,0.01900800069173177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,768,0.07940977811813354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,10240,1.6958008872138128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,16384,0.10757599936591254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,1536,0.055000000529819064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,16384,0.6594124370151097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,6144,0.03191466795073615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,1536,0.16124355792999268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,6144,0.022008889251285132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,256,0.004570666700601578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,51200,0.5053155687120225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,12288,0.017419555120997958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,1024,0.07047377692328559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,12288,0.37494577301873105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,6144,0.012864000267452665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,10240,0.02648799949222141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,1024,0.020251555575264823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,3584,0.006737777756320105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,1024,0.0041582224269707995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,768,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,12288,0.4624684651692708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,6144,0.2818888823191325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,768,0.004879999905824661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,1024,0.009848888549539778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,5120,0.10545777612262303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,256,0.00906755526860555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,16384,0.029641777276992798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,64,0.005212444398138259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,6144,0.012615111139085559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,768,0.00573600000805325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,3072,0.061420447296566434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,7168,1.0050151613023546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,2560,0.007696000238259633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,8192,0.038557334078682795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,2560,0.009560888840092553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,64,0.0029804443733559716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,128,0.0027066665805048416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,16384,1.271685282389323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,768,0.04457777738571167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,64,0.007315555380450354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,4096,0.008850666383902231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,32,0.003565333369705412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,1024,0.006807111203670502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,128,0.004218666503826777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,64,0.010850666297806634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,512,0.004039999925427967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,3072,0.03772888912094964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,32,0.011481777661376528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,2048,0.0074035558435651995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,1024,0.028219554159376357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,32,0.06100889046986898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,7168,0.46615558200412327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,8192,0.015972443752818637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,7168,0.23240354326036242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,3072,0.006313777632183499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,16384,0.01885333326127794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,1024,0.006903111106819577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,3584,0.05087200138303968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,16384,0.6149395306905111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,3584,0.10077155298656887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,2560,0.022459555003378127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,3584,0.006871999965773688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,64,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,512,0.19308355119493273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,256,0.00407288885778851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,768,0.004275555411974589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,7168,0.015511110424995422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,6144,0.14328444004058838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,64,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,256,0.008192888564533657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,64,0.005331555588377847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,768,0.006193777753247156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,128,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,3072,0.008062221937709385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,51200,0.07589155435562134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,7168,0.13252445062001547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,1536,0.014025777578353882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,2560,0.01533600025706821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,5120,0.047129776742723256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,51200,1.6209946738349066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,65536,2.4615190294053817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,51200,2.949817869398329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,3072,0.029882666137483384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,7168,0.05151199963357714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,12288,0.28424178229437935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,51200,0.35115912225511337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,64,0.0030622221529483795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,2560,0.03487644592920939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,3072,0.033040000332726374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,2048,0.6509315702650282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,1024,0.0064106666379504735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,2048,0.060984002219306103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,768,0.08603377474678887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,3072,0.009519110951158736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,32,0.011016888750924004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,2048,0.013383111192120446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,32,0.03641511003176371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,16384,0.021418665846188862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,256,0.0029680000411139596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,2048,0.007135999699433644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,256,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,7168,1.0788372887505424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,16384,0.31832533412509495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,3584,0.010223111344708337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,128,0.0053191110491752625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,3584,0.010682666467295753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,3584,0.010486221975750394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,10240,3.06189325120714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,128,0.009099555512269339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,1024,0.00776711106300354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,2048,0.03313688768280877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,768,0.06254311402638753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,3584,0.033118221494886614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,4096,0.024354666471481323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,4096,0.20492533842722574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,51200,0.1805457803938124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,65536,6.435406155056423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,10240,0.02696177860101064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,4096,0.009975999593734741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,4096,0.014816888504558139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,51200,0.01644622286160787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,128,0.01607733302646213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,256,0.01649777756796943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,4096,0.015740444262822468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,8192,0.015231110983424716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,1536,0.00886933339966668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,128,0.004054222255945206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,51200,0.13812266455756292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,32,0.0027742222365405825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,1024,0.003949333396222857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,16384,3.4848738776312937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,32,0.0035822221802340615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,4096,0.36024265819125706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,6144,0.006508444332414203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,51200,0.2162302202648587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,6144,0.019428445233239066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,8192,3.333594640096029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,1536,0.010544000400437249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,8192,0.022893332772784766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,256,0.0038542221817705366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,10240,0.01790577835506863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,3072,0.016389333539538913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,65536,0.9955875608656141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,12288,0.5167537795172797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,512,0.006276444428496891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,3584,0.01126399967405531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,12288,0.03840000099605984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,8192,0.4220435619354248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,2048,0.8238444328308105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,5120,0.021200888686709937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,1536,0.005897777775923411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,3072,0.05349777804480659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,16384,0.061831997500525586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,12288,0.3754711151123047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,5120,0.11441689067416722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,256,0.005774222314357758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,256,0.012295111185974546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,8192,0.2342382272084554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,256,0.005264889034960005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,3072,0.007054222126801808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,3072,0.026706667410002813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,6144,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,7168,0.022443556123309668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,7168,0.011425777441925473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,16384,0.1440204381942749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,65536,0.43203287654452854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,2560,0.008222222328186035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,2048,0.010119111173682744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,768,0.003714666598372989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,6144,0.008584000170230865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,51200,0.6898630989922417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,768,0.03821333249409994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,64,0.00407911092042923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,32,0.003064888839920362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,65536,0.2131342225604587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,12288,0.06323111057281494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,65536,0.9071600172254773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,16384,0.40575644705030656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,1536,0.006299555715587404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,5120,0.013799111048380533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,768,0.004397333496146732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,10240,0.2249057822757297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,3584,0.00867555538813273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,128,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,1536,0.007335999773608313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,256,0.0069013333155049225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,256,0.003101333354910215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,10240,0.300105783674452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,8192,0.012282667060693106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,16384,0.4287422233157688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,2560,0.08449066347546047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,32,0.002510222295920054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,65536,2.935010698106554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,3584,0.8614302211337619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,12288,0.024021334118313257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,256,0.007639110916190677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,128,0.0032799999333090256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,2048,0.015119110544522604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,3584,0.021496888664033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,768,0.005454222361246745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,12288,0.05810578001870049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,256,0.021018667353524104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,51200,0.028852442900339764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,6144,0.032644444041781955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,4096,0.02792533238728841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,3584,0.04714399907324049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,2048,0.024616888827747766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,12288,0.07867022355397542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,2560,0.0687991115781996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,12288,5.983287811279297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,32,0.0069217777086628815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,2560,0.01869777838389079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,64,0.006737777756320105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,6144,0.09317600064807469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,8192,0.0528035561243693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,256,0.012135110795497894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,3584,0.010510221951537661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,256,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,128,0.09027022123336792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,12288,0.08047822448942396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,512,0.030615111192067463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,768,0.009591999981138442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,1536,0.010565333068370819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,2048,0.023191110955344305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,256,0.0030577776746617425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,12288,0.021534222695562575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,16384,0.06587022542953491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,3072,0.18505777253044975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,4096,0.015014222926563688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,51200,0.019181332654423185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,2560,0.016576889488432143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,1536,0.032489776611328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,10240,0.012998222476906247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,1536,0.01946311030122969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,3072,0.14896889527638754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,32,0.008867555194430882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,2560,0.26300711101955837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,51200,0.28027910656399196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,1536,0.14175466696421304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,7168,0.04984177814589607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,10240,0.4254471196068658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,5120,0.30981598960028756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,5120,0.08412800232569377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,2560,0.057074666023254395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,768,0.047269334395726524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,2560,0.009922666682137383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,3584,0.07008888986375597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,2560,0.2715448803371853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,16384,0.17357867293887666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,1536,0.016583111551072862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,1024,0.004366222355100843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,12288,16.185697767469616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,10240,0.03689422210057577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,2048,0.00908799966176351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,8192,0.4035164515177409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,12288,0.020492444435755413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,2048,0.007275555696752336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,4096,1.6040062374538844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,7168,0.20663911766476104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,5120,0.013611555927329592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,3584,0.009895111123720804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,65536,1.0999013053046334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,1024,0.026294221480687458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,1536,0.005797333187527127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,2048,0.007113777928882175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,1024,0.036122666464911565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,512,0.017224889662530687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,5120,0.009882666998439366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,3584,0.014703111516104804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,3072,0.008256889051861232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,3072,0.8828693495856391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,256,0.006981333096822103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,3072,0.009151111046473185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,7168,0.4868577851189508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,8192,0.9860489103529188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,128,0.020091555184788175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,512,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,10240,0.01633066601223416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,65536,1.5140000449286566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,5120,0.01201866649919086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,7168,0.02660622199376424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,4096,0.015501333607567681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,65536,0.24894576602511934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,1024,0.011153777440388998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,1536,0.08649510807461208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,3584,0.16719378365410698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,7168,0.2098942200342814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,1024,0.005056000004212062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,32,0.005241777747869492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,256,0.004598222259018156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,4096,0.012611555556456247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,64,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,5120,0.01481244464715322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,3584,0.26883554458618164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,128,0.10211288928985596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,1536,0.006296000132958095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,6144,0.021327111456129286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,51200,0.05043199989530775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,3072,1.4290453592936199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,128,0.02477866742346022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,256,0.0034248888906505373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,16384,0.1658115519417657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,5120,0.008815999660227034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,1536,0.05419466561741299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,32,0.002598222138153182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,1024,0.006270222365856171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,3584,0.10841777589586045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,768,0.008360000120268928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,768,0.006237333433495627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,3584,0.048229333427217275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,10240,0.006225777582989798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,16384,0.029860443539089624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,7168,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,3584,0.01316088851955202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,3072,0.00925244473748737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,3584,0.007732444339328342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,7168,0.03664355476697286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,768,0.003657777690225177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,6144,0.019697777099079557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,768,0.007621333003044128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,10240,0.02757599949836731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,3072,0.010970667004585266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,12288,0.43576712078518337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,1536,0.016403555870056152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,16384,0.01806400054030948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,8192,0.06304977999793158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,65536,0.0949831141365899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,10240,0.029080000188615587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,256,0.004726222405831019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,3584,0.1110337773958842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,6144,0.012288889123333825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,1024,0.008085333638721043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,16384,0.04046488801638285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,2048,0.0057964444988303715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,64,0.02769600020514594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,65536,0.037639998727374606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,5120,0.017633777525689866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,1536,0.025281776984532673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,1024,0.010856000085671743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,128,0.0038844446341196695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,256,0.013002666334311167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,5120,0.012778666284349231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,4096,0.02100444502300686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,2560,0.17806222703721789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,8192,0.2602035469479031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,2048,0.011047110789351992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,1024,0.006787555499209299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,64,0.0025724443710512584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,3072,0.10192710823482937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,128,0.005031110925806893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,7168,0.3192444377475315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,6144,0.19549422793918184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,5120,0.013108444710572561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,1024,0.016023111012246873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,3584,0.012758221891191272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,2560,0.006119110931952794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,65536,5.7816649542914496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,1536,0.07108533382415771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,8192,0.03039466672473484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,65536,0.19414844777848986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,16384,0.15680444240570068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,4096,0.0310204459561242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,32,0.007174222005738153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,12288,0.08594044711854722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,10240,0.6509422196282281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,12288,0.058803558349609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,3584,0.010791999598344168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,32,0.006558222075303395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,5120,0.03497688968976339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,7168,0.1798631085289849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,3072,0.007888889147175683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,768,0.015542222393883599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,128,0.004400889078776042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,768,0.064282668961419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,768,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,65536,2.2519520653618708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,32,0.01331911070479287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,5120,0.019920888874265883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,2560,0.012030222349696688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,1024,0.10619199938244289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,3072,0.016082665986484952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,65536,2.2043599022759333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,6144,0.2782284418741862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,10240,0.0343831115298801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,3072,0.8692302174038358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,4096,0.0861626664797465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,51200,0.06077510780758328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,6144,0.0068862222962909276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,10240,0.102692445119222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,32,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,1024,0.009465778039561378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,3072,0.011146666275130378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,16384,0.019275556008021038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,51200,0.23177689976162383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,1536,0.14099289311303034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,8192,0.019103111492262948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,3584,0.01071999967098236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,5120,0.33272798856099445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,512,0.006715555571847492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,256,0.008356444537639618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,512,0.007496889266702864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,3072,0.09272799889246623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,32,0.008433777425024245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,12288,0.02751200066672431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,128,0.0031804444475306403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,3072,0.014214222629865011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,2560,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,3072,0.5025368796454536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,65536,0.25596621301439076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,2048,0.23812799983554414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,8192,0.058789332707722984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,4096,0.657275570763482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,51200,0.7674764527214898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,64,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,3072,0.0075528886583116316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,8192,0.21818933222028944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,5120,0.008808888494968414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,512,0.026890666948424444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,12288,0.0686053302552965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,512,0.005371555685997009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,32,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,32,0.011140444212489657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,12288,0.020444444484180875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,51200,0.2427288956112332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,3072,0.009580444130632613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,768,0.005766222046481238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,1024,0.009136888715955947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,16384,0.22258488337198892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,6144,0.05805066559049818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,10240,0.29081688986884224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,65536,0.06925955745908949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,1536,0.05327733357747396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,7168,0.027495111028353374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,10240,0.3486248917049832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,512,0.005550222264395819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,3072,0.12428889009687637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,768,0.021503110726674397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,1024,0.006992888947327931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,65536,0.29490754339430064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,64,0.0052062223354975385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,2560,0.007096000015735626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,12288,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,65536,0.23407554626464844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,512,0.004543999830881755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,16384,0.4238479932149251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,7168,0.016975111431545682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,8192,0.04719733198483785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,3072,0.05704533391528659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,12288,0.0861733357111613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,256,0.01443377799457974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,2048,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,8192,0.021918222308158875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,1024,0.37111555205451113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,7168,0.3705288834042019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,256,0.007527111305130853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,7168,0.03132355544302199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,256,0.004449777719047335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,6144,0.011367110742463006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,2560,0.010653333531485664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,256,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,1536,0.025983111725913152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,6144,0.01391377713945177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,8192,0.9669378068712023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,10240,0.008759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,512,0.06552355819278292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,2560,0.0140684445699056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,2048,0.008732444710201686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,2560,0.01129066695769628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,51200,0.6155502001444498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,8192,0.012696888711717395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,256,0.006248888870080312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,8192,0.02941511074701945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,128,0.008867555194430882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,12288,0.08328000042173597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,2048,0.006752000086837345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,4096,2.762389289008247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,65536,2.918886184692383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,2560,0.008181333541870117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,64,0.003003555453485913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,128,0.0034168888297345904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,16384,0.6771981981065539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,2048,0.07012177838219537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,51200,0.24168888727823892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,51200,0.21948711077372232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,128,0.0029964444951878656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,3072,0.2503431108262804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,65536,0.013075555364290873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,1536,0.03695111142264472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,64,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,16384,0.05019022358788384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,7168,0.05135022269354927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,2048,0.015847111741701763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,2048,0.022268444299697876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,65536,0.42408090167575413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,16384,0.04237511091762119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,3584,0.010943111446168689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,8192,0.023151111271646287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,2560,0.044396443499459155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,65536,2.421760982937283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,3584,0.041486223538716636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,256,0.0031439999325407874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,1024,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,51200,0.1365333398183187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,4096,0.02734311090575324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,3584,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,51200,0.00887288898229599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,64,0.0026986667265494666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,10240,0.015990222493807476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,4096,0.32134665383232963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,3584,0.5079413519965278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,5120,0.01815200017558204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,1024,0.017855111095640395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,128,0.00498488876554701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,3072,0.013095999757448832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,512,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,128,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,5120,0.016571554872724745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,512,0.007804444266690149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,8192,0.03173689047495524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,32,0.011975111232863532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,2048,0.01073155552148819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,512,0.03121688961982727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,7168,0.0384897788365682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,4096,0.008811555802822113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,512,0.006816888848940532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,256,0.005943111247486538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,3072,0.0517813331551022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,128,0.004060444318585926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,12288,0.6827831268310547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,2048,0.009679111341635386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,1024,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,5120,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,5120,0.033360888560613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,51200,0.5095538033379449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,8192,0.041313777367273964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,768,0.11701867315504287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,65536,1.4622284571329753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,3584,0.1085404422548082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,10240,0.5276355743408203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,2048,0.007920888562997183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,2560,0.010827555424637265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,4096,0.012097777591811286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,65536,0.06378133429421319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,65536,0.023208000593715247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,10240,0.10493333472145928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,10240,0.8034328884548612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,1536,0.04290133383538988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,4096,0.10726488961113824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,12288,0.033634665939542986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,12288,0.43292978074815536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,32,0.008373333348168267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,512,0.005392888767851724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,10240,0.04849066668086582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,65536,0.14014400376213923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,1536,0.006777777853939269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,10240,0.2032888862821791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,4096,0.01701955497264862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,512,0.003354666754603386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,16384,3.585657755533854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,5120,0.016657777958446078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,16384,0.22229332394070098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,5120,0.0517173343234592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,4096,0.011022222538789114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,128,0.0036524443162812125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,8192,0.11416888236999512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,8192,0.01162755572133594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,65536,0.7178719838460287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,8192,0.024304888314670984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,12288,0.03468799922201369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,768,0.004679999831649992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,3072,0.021537777450349595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,16384,0.06399910979800753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,2560,0.015375999940766228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,512,0.007155555817815993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,16384,0.04981599913703071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,51200,2.372384813096788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,768,0.009230222139093611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,256,0.004407999830113517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,7168,0.015480889214409722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,8192,0.021015110943052504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,16384,0.22243110338846842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,8192,0.031160887744691636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,1024,0.004035555654101902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,3072,0.2620675563812256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,51200,3.208014170328776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,128,0.008746667040718926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,768,0.005727111051479976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,6144,0.16564355956183538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,51200,0.008442666795518663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,1024,0.004534222185611725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,64,0.004635555462704765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,3584,0.021407110823525324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,65536,0.2824409008026123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,16384,1.9902559916178386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,65536,0.27205244700113934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,10240,0.0892088876830207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,6144,0.009723555710580613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,51200,0.32450487878587514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,2048,0.009000000026490953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,32,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,16384,0.03120888935195075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,8192,0.17829067177242705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,4096,0.010643555886215635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,256,0.004572444491916233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,1024,0.01052266690466139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,4096,0.011039999624093374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,32,0.010289777484205034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,65536,0.18163377708858916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,768,0.0255404445860121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,2048,0.013807111316257052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,1024,0.005008000052637524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,4096,0.06360177861319648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,4096,0.014256000518798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,1536,0.010630221830474006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,3584,0.04627111223008898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,10240,2.1155367957221136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,1536,0.06805688805050321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,5120,0.16720978418986002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,32,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,1536,0.00871733327706655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,256,0.00850399997499254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,8192,0.16724088456895617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,6144,0.13263111644321018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,16384,0.40060088369581437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,4096,1.0425643920898438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,3072,0.08910577827029759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,12288,0.05009066727426317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,512,0.00406577769252989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,65536,2.6778471204969616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,1024,0.00777511133088006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,7168,0.011980444192886353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,4096,0.016116445263226826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,51200,22.55221388075087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,4096,0.013094222380055321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,128,0.005741333381997214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,65536,0.10477066702312893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,65536,1.6364755630493164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,128,0.006149333384301927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,16384,0.05981777773963081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,10240,2.094413333468967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,8192,3.7956131829155813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,3072,0.010769777827792697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,2560,0.009205333060688442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,32,0.0034044444974925783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,8192,0.015311999453438653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,3072,0.08538666698667739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,1536,0.28817065556844074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,1536,0.005752000129885144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,2048,0.008741333252853816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,1536,0.007283555964628856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,16384,0.39218934377034503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,256,0.0040960001448790235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,6144,0.044307556417253285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,6144,0.025855110751258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,128,0.012807110945383707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,10240,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,1536,0.015263999501864115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,65536,1.1628292931450737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,12288,0.01930488811598884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,12288,0.10529688994089763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,1536,0.038358221451441445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,12288,0.016012445092201233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,2048,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,128,0.013767999907334646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,16384,0.483717335595025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,4096,0.08174666431215075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,5120,1.4032871458265517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,12288,0.047487109899520874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,1536,0.02201066745652093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,7168,0.018352889352374606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,256,0.004810666872395409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,2560,0.0509422222773234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,51200,0.4857609007093642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,128,0.005505777895450592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,65536,8.536133660210504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,8192,0.10860355695088704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,768,0.007721777591440413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,5120,0.021764443980322942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,4096,0.011703111231327057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,6144,0.024639111426141527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,5120,0.010568000376224518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,512,0.006226666685607698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,512,0.008454221818182204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,1024,0.005112888912359874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,64,0.005230222311284807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,8192,0.007316444483068254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,51200,0.09771022531721328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,2048,0.009751111268997192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,6144,0.022252443763944838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,10240,0.013340444200568728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,64,0.00507377791735861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,16384,1.3242257436116536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,128,0.009309333231714036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,5120,2.8446862962510853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,64,0.053519109884897866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,768,0.007551111280918121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,3072,0.009749333063761393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,8192,0.09831288788053726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,6144,0.015448000695970325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,1024,0.04251911242802938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,3584,0.010885333021481832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,256,0.023390221926901076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,4096,0.04154310954941644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,65536,1.2391306559244792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,1536,0.34248889817131895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,12288,0.21768532858954537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,16384,0.033111111985312566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,64,0.007596444752481248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,2048,0.031680001152886286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,1024,0.004358222087224324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,3072,0.05566844675276014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,1024,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,16384,0.007655111451943715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,2048,0.06441689199871488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,256,0.008176888856622908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,128,0.0026213334252436957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,512,0.06640266709857516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,7168,0.09704711039861043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,2048,0.04404888881577385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,65536,0.07834311326344807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,12288,0.0181697772608863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,512,0.004328888737493091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,2560,0.02644266684850057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,1536,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,10240,0.026743110683229234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,3584,0.01687022215790219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,65536,2.4666275448269315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,1024,0.007345777418878343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,2560,0.008956444760163626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,3072,0.015480889214409722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,65536,0.2985315587785509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,128,0.013230221966902414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,6144,0.010644444160991244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,3072,0.023296000228987798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,3072,21.099356757269966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,6144,0.07117511166466607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,1536,0.01274577776590983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,1536,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,4096,0.07679555813471477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,16384,0.022523555490705702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,768,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,256,0.0034977777136696708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,10240,0.23488799730936685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,65536,0.6829475296868218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,2560,0.025974222355418738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,8192,0.13817511664496526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,51200,1.2849431567721896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,65536,3.0008985731336804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,3584,0.021432888176706102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,65536,0.07900800307591756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,10240,0.07922844754325019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,8192,0.024327110913064744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,256,0.005296000176005893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,7168,0.09176977475484212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,12288,0.02903466754489475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,32,0.0042364444169733255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,4096,0.006550222221348021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,6144,0.017398221625222098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,64,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,12288,0.04647466540336609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,5120,0.01566488875283135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,2560,0.010852444503042432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,16384,0.19860355059305826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,1024,0.004372444417741564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,16384,0.6257866753472222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,768,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,1024,0.006879111131032308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,32,0.002768888862596618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,5120,0.19960533248053658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,6144,0.023407111565272015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,768,0.009114666945404476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,256,0.043395555681652494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,8192,0.02536533276240031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,65536,2.342967987060547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,8192,0.020459555917316012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,12288,0.19165155622694227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,2560,0.0785964462492201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,7168,0.5382640096876357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,4096,0.011986666255527072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,256,0.008943999807039896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,3584,1.7337208853827581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,256,0.00583022211988767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,65536,0.6971306800842285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,10240,0.15973156028323704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,64,0.010244444840484196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,65536,0.012351999680201212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,2048,0.008396444221337637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,256,0.008966222405433655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,12288,0.022413333257039387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,1024,0.009409777820110321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,6144,0.8784586588541666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,4096,0.008184889124499427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,8192,0.11262222131093343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,2560,0.12193955315483941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,6144,0.016533333394262526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,2560,0.011094222466150919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,8192,0.028981334633297388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,5120,0.008734222087595198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,3072,1.5241964128282335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,5120,0.04241066508822971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,1024,0.037893331713146634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,512,0.004848888764778773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,256,0.0048177776237328844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,256,0.005507555686765247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,16384,0.4421786732143826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,10240,8.863455878363714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,65536,3.832533518473307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,10240,0.5699546602037218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,16384,0.022074666288163926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,256,0.006301333506902059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,64,0.005694222119119432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,128,0.004664888812435998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,16384,0.05512711074617174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,128,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,6144,0.10758755604426067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,2560,0.02323466704951392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,65536,0.8342649141947428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,32,0.0040995557275083335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,10240,5.2242452833387585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,6144,0.024359110328886244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,1536,0.007725333174069722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,5120,0.017827555537223816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,768,0.24023734198676217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,16384,0.026690666874249775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,3072,0.028925332758161757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,768,0.0037502224246660867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,16384,0.5481724209255642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,3584,0.1178026729159885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,1024,0.00702133360836241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,4096,0.008626666333940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,64,0.0029680000411139596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,7168,0.017690666847758822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,768,0.01808888879087236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,3072,0.0800515545739068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,3072,0.008915555146005418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,256,0.006067555397748947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,768,0.006236444330877728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,5120,0.009335999687512716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,10240,0.04385066694683499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,2560,0.006075555665625467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,1536,0.032990223831600614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,512,0.004057777838574516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,768,0.02745955520206028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,16384,0.025060445070266724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,8192,0.07438044415579902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,128,0.008663111262851292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,3584,0.007910221815109253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,1536,0.0595466693242391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,3072,0.2616159915924072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,10240,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,2560,0.0068195557428730865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,51200,0.06052977508968777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,2560,0.011357333097192975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,128,0.006598222172922558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,1536,0.014711111783981323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,51200,0.06555022133721246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,1536,0.006454222318198945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,128,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,256,0.014645333091417948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,768,0.009381333159075843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,512,0.007327111230956183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,4096,0.07389422257741292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,3584,0.012016889121797351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,6144,1.3424338234795465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,5120,0.01644088824590047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,768,0.007891555627187094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,32,0.003346666693687439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,51200,3.0216382344563804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,4096,0.39052711592780215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,1024,0.040022220876481794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,512,0.012730666332774691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,6144,0.030716445710923936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,2560,0.04119022356139289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,6144,0.01272266689274046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,4096,0.011977777712874942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,768,0.09281955824957953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,1024,0.005200888961553574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,51200,0.7195866902669271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,2048,0.010716444088353051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,10240,0.1458915604485406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,2048,0.012914666699038612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,1024,0.0418488879998525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,5120,0.02032444377740224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,10240,0.43391021092732746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,6144,0.038895110289255776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,12288,0.03124799993303087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,1536,0.014819555812411837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,2048,0.00998755544424057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,768,0.008016889293988546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,6144,1.0128747092352974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,1024,0.011010666688283285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,64,0.00610133343272739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,3584,0.010574222438865237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,6144,0.014524444937705994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,128,0.0030826667530669104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,12288,0.01983022193113963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,1024,0.006627555522653792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,256,0.00977955593003167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,5120,0.02852533260981242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,256,0.003630222131808599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,2048,0.09155378076765273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,10240,0.7236008644104004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,768,0.028400888045628864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,1024,0.00776800016562144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,7168,0.019139554765489366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,3072,0.008469333251317343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,3584,0.023600889576805964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,16384,0.0273297776778539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,7168,0.018550222118695576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,10240,0.367470211452908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,768,0.003965333518054751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,1536,0.0063946665161185795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,1024,0.008336889247099558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,7168,0.04718044400215149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,256,0.004734222259786394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,3072,0.16262400150299072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,512,0.05406844284799364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,512,0.0031635556370019913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,768,0.0057955553962124725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,4096,0.021310221817758348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,1024,0.07118844323688082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,2048,0.009977777798970541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,1024,0.004039111236731212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,65536,0.17218844095865884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,7168,0.016182222300105624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,768,0.030810667408837214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,7168,1.7594310972425673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,12288,0.2515590985616048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,128,0.003548444559176763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,2048,0.004986666556861666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,6144,0.08103377951516046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,1024,0.0355751117070516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,256,0.03015822172164917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,65536,0.15016710758209229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,2048,0.2548533280690511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,4096,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,2560,0.11219288243187799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,4096,0.023904000719388325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,512,0.33500533633761936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,10240,0.12354489167531331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,5120,0.028312000963422988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,1536,0.004840888910823398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,768,0.008450667063395182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,65536,0.48854578865898984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,6144,0.012863111164834766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,6144,0.010568888651000129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,5120,0.12779555055830213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,64,0.00699733363257514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,12288,0.02019999921321869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,2048,0.04440800017780728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,2048,0.006711111300521427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,3584,0.008169777691364288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,8192,0.06627911329269409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,4096,1.1112417644924588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,5120,0.013559111290507846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,5120,0.031165334913465712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,32,0.05804710918002658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,16384,0.08595644765430027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,8192,0.6691999965243869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,512,0.006671999891599019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,51200,12.462877909342447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,1536,0.013049778011110095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,128,0.05199644300672743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,10240,0.02295111119747162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,512,0.00848444468445248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,12288,0.4684986538357205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,32,0.006890666567616993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,512,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,1024,0.006314666734801398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,3072,0.010346666806273991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,6144,0.017647110753589206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,5120,0.2689555486043294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,10240,0.0411102208826277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,8192,0.012725333372751871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,128,0.016527111331621807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,3584,0.03773066732618544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,65536,0.14530489179823133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,16384,0.031152000029881794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,6144,0.012913777596420713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,128,0.007868444754017724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,3584,0.3084239959716797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,3584,0.05372888843218485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,64,0.007252444823582967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,64,0.007347555624114142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,128,0.022711111439598933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,2560,0.006146666490369373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,3584,0.22884265581766763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,3072,0.011392888923486074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,4096,0.07306577761967976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,5120,0.050832887490590416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,256,0.0032479998966058097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,7168,0.023260444402694702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,1536,0.04395022326045566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,7168,0.026906667484177485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,512,0.014975999792416891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,3584,0.06383466720581055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,7168,0.017811555001470778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,6144,0.009578666753239101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,768,0.32968176735772026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,7168,0.010330666270520952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,4096,0.06674222151438396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,51200,1.6828781763712566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,4096,0.07102044423421223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,6144,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,7168,0.2035155561235216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,51200,0.7170453071594238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,64,0.0034755555291970572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,8192,0.06769600179460314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,256,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,6144,0.3027893437279595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,10240,0.15229422516292995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,12288,0.6347635587056478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,2560,0.02183466653029124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,10240,0.0795288880666097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,256,0.01625955601533254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,2560,0.007752888732486301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,6144,0.03589955634540982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,768,0.00462666650613149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,10240,0.18437600135803223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,5120,5.2490285237630205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,5120,0.025557332568698462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,12288,0.4633333418104384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,2560,0.009293333523803288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,10240,0.45522223578559023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,10240,0.35430222087436253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,6144,0.09790400001737808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,768,0.02830044428507487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,2048,0.009469333622190688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,5120,0.05652799871232775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,2048,0.009842666486899057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,2048,0.008311111066076491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,5120,0.007666666474607255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,2048,0.07157333029641046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,65536,0.16423022747039795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,51200,0.26008444362216526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,2048,0.012885332935386233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,7168,0.8329226705763074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,2048,0.010973333484596677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,2048,0.03423377871513367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,6144,0.1632631089952257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,7168,0.5418835745917426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,16384,0.09828355577256943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,1536,0.010932444698280759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,32,0.004664888812435998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,3584,0.4924489127265082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,8192,0.017094223035706412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,7168,0.04447199900945028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,65536,0.08927466471989949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,768,0.005486222190989389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,2560,0.01635199950801002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,64,0.0026142222599850762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,1024,0.007920888562997183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,65536,0.35658934381273055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,128,0.009670221971140968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,512,0.013278221918476952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,64,0.00397155558069547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,7168,0.19410933388604057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,2560,0.0063733334342638654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,6144,0.02254755629433526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,16384,0.4450497627258301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,1024,0.021548445026079815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,7168,0.8769484625922309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,5120,0.08309955729378594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,2560,0.12084710597991943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,2560,0.007155555817815993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,51200,5.7880443996853295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,1536,0.027375110321574744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,6144,0.2876053386264377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,12288,0.022239999638663396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,12288,0.022120889690187242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,1024,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,128,0.0038684445122877755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,10240,6.368918948703342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,2048,0.08349777592553033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,2560,0.01293155550956726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,5120,0.014045332868893942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,65536,0.4153297742207845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,16384,0.03742311067051358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,6144,0.054768890142440796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,5120,0.016864000095261466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,65536,8.205476548936632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,8192,0.011146666275130378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,16384,0.5561448733011881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,512,0.02753422160943349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,1024,0.006696888970004187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,768,0.022983110613293115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,768,0.006026666611433029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,8192,0.15062488449944392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,3072,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,32,0.11963289313846165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,10240,0.041823112302356295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,12288,0.12279733022054036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,7168,0.014369777507252164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,8192,0.07010666529337566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,2048,0.16802222198910186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,64,0.0027502222607533136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,10240,0.017272889614105225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,256,0.006896889044178857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,7168,0.14837333891126844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,512,0.02383288906680213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,768,0.008810666700204214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,1536,0.011665778027640449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,8192,0.7079786724514432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,12288,0.3323635525173611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,64,0.0027724444452259275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,12288,0.578146669599745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,3072,1.273158179389106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,51200,0.072561780611674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,2048,0.015652444627549913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,5120,0.013688888814714221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,768,0.015095111396577625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,768,0.24417956670125326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,768,0.014546665880415173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,768,0.02027555637889438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,128,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,3584,0.01074488874938753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,512,0.003567111161020067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,3584,0.014536889062987434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,768,0.020614221692085266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,8192,0.27320978376600474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,3072,0.1608497831556532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,64,0.005316444569163852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,512,0.011848889291286469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,12288,0.015187554889255099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,32,0.0027155555370781156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,256,0.008749333520730337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,64,0.17372355196211073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,6144,0.038952890369627215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,12288,0.019354666272799175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,8192,0.018000000052981906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,1024,0.025978666212823655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,2048,0.02836266822285122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,1024,0.014465777410401238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,128,0.011969777444998423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,768,0.033088889386918806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,5120,0.012885332935386233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,51200,12.282318115234375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,32,0.006695999867386288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,1536,5.768032921685113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,256,0.02932888931698269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,64,0.005967111223273807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,32,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,8192,0.03149155444569058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,51200,0.12133688396877712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,512,0.026703110999531213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,1536,0.008986666798591614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,3072,0.03443822264671326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,51200,0.09325333436330159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,1536,0.011916444533401065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,768,0.10506221983167861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,65536,0.10846400260925293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,10240,1.2902666727701824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,51200,0.1517502201928033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,768,0.00599644457300504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,3584,0.019914666811625164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,16384,0.04511022236612108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,768,0.017640888690948486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,1024,0.029686222473780315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,12288,0.06628444459703234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,2048,0.024455999334653217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,16384,0.08347555663850571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,32,0.005836444596449534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,256,0.022338666849666174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,1024,0.008728000024954477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,6144,0.010468444062603844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,32,0.0340942210621304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,3072,0.009773333039548662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,3072,0.017208000024159748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,2560,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,512,0.003972444269392225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,128,0.0162444445821974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,4096,1.2933190663655598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,3584,0.05157422357135349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,3584,0.11841421657138401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,6144,0.09718933370378281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,51200,0.24047999911838105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,2560,0.009560888840092553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,2560,0.19780266284942627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,5120,0.38151733080546063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,16384,0.1715893348058065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,3072,0.22363999154832628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,7168,0.031776888502968684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,8192,0.015100444356600443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,5120,0.017802667286660936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,16384,0.3452177842458089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,128,0.004593777573770947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,128,0.006020444548792309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,10240,0.23785066604614258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,5120,0.1067911121580336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,3072,0.07616177532407972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,5120,0.06387288702858819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,4096,0.05929155482186211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,1024,0.010037333601050908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,1536,0.09847288661532932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,4096,0.09216088718838161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,8192,0.5439324378967285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,10240,0.18781867292192247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,64,0.0029893333299292457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,12288,0.03271377748913235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,128,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,65536,0.06971555285983615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,4096,0.014166222678290473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,2048,0.08148444361156888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,4096,0.008234666453467475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,768,0.007794666621420119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,3072,0.3899537722269694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,1536,0.0473582214779324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,64,0.013425778183672162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,12288,0.19295556015438506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,65536,0.09269244141048855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,12288,0.08720978101094563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,3584,0.0184515549076928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,8192,0.05256266726387871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,12288,0.01000711073478063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,2048,0.06258489025963677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,65536,0.01815555493036906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,10240,0.09386489126417373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,5120,0.017255999975734286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,3072,0.10436444150076972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,10240,0.013377777404255338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,16384,0.15113243791792128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,12288,0.19179733594258627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,64,0.0053484443989064955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,128,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,3584,0.14749688572353786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,12288,0.31921421156989205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,7168,0.030759112702475652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,2560,0.006881777611043718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,51200,0.5194391144646539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,3584,0.016584888100624084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,512,0.005036444299750858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,2048,0.007285333342022366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,6144,0.177985774146186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,10240,0.027181333965725366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,128,0.006814222368929122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,4096,0.1823155615064833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,65536,0.3838675551944309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,10240,0.05414666732152303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,1024,0.00674222202764617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,65536,1.7463999854193792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,4096,0.015408888459205627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,2560,0.008059555457697975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,2560,0.006026666611433029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,64,0.008504889077610439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,128,0.008242666721343994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,51200,0.11621510982513428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,512,0.004351110921965705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,512,0.04352533486154345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,128,0.007126222054163615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,768,0.005821333163314396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,8192,0.08446577522489761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,3072,0.015885333220163982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,64,0.21351644727918836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,1536,0.029808001385794744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,3584,0.020971554848882887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,2560,0.015088000231319003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,128,0.0026906666656335196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,3584,0.006719999843173557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,512,0.004514666481150521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,51200,0.09628622399436103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,3584,0.15111200014750162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,1536,0.006855999843941794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,1024,0.0047484445903036325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,65536,0.3705893357594808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,2048,0.06639822324117024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,16384,0.1013759970664978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,1536,0.007285333342022366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,8192,0.2138079934649997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,65536,1.7509280310736761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,6144,5.709537929958767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,65536,23.999571058485245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,16384,0.024223110742039148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,1536,0.05588533480962118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,1536,0.03397599856058756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,3072,0.025006221400366888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,6144,0.017330666383107502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,1536,0.005576888720194499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,512,0.003929777691761653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,5120,0.19419288635253906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,12288,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,7168,0.02665422194533878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,512,0.004838222016890843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,16384,0.03989333245489333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,768,0.008663111262851292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,768,0.004908444566859139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,16384,0.025869333081775244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,128,0.0053635554181204895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,7168,0.02806666824552748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,32,0.0038115556041399636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,2560,0.008897778060701158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,256,0.0037457777394188773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,65536,0.07033244768778484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,128,0.004585777719815572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,6144,0.11921066708034939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,8192,0.12111111481984456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,1536,0.4733155568440755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,8192,0.021670222282409668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,1536,0.014879999889267815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,1024,0.005375111268626318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,5120,0.15033066272735596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,3584,0.025437333517604407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,10240,0.025736888249715168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,32,0.002628444383541743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,10240,0.05029777685801188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,3584,0.016249777542220224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,10240,0.0071937781241205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,7168,0.01907911068863339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,6144,0.010656000011497073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,65536,3.59623294406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,7168,0.13076799445682102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,1536,0.019925332731670804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,3072,0.010155555274751451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,8192,0.12507732709248862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,8192,0.01571555601225959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,3072,0.008050666915045844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,512,0.010775110787815519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,1536,0.010008888940016428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,768,0.003522666792074839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,2048,0.07049777772691515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,3584,0.01740711099571652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,512,0.004150222159094281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,65536,0.6581404474046495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,2048,0.030576888057920668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,128,0.009395555489593083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,4096,0.012119111087587146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,6144,0.473583115471734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,12288,5.109052446153429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,1536,0.01979733341270023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,5120,0.09813955757353042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,5120,0.013292444248994192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,7168,0.021007999777793884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,16384,0.14558666282229954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,10240,0.5195493168301052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,8192,0.03469333383772109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,128,0.002983999955985281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,6144,0.4043573273552789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,5120,0.04822488956981235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,16384,0.03424533208211263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,3072,0.10328355762693618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,16384,0.20104977819654676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,3072,0.41590044233534074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,32,0.0033031110134389666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,8192,0.01531644496652815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,10240,0.09361600213580662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,51200,0.5670951207478842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,512,0.010637332995732626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,256,0.005467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,8192,0.26240799162122935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,3072,0.11083555883831447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,2048,0.053415109713872276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,2048,0.010243555737866296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,256,0.005671111245950063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,65536,0.23160089386834037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,12288,0.3293626573350695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,16384,0.2052391105228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,6144,0.044454223579830594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,256,0.016590222716331482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,12288,0.015049777097172208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,65536,1.4973618189493816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,32,0.00434933313065105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,3584,0.021480000681347314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,51200,3.092100355360243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,64,0.0032239999208185407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,2560,0.11771199438307021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,8192,0.04260622130499946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,2048,0.05603288941913181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,1536,0.005184000151024925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,7168,0.011352889239788055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,12288,0.12829244136810303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,4096,0.0358951124880049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,7168,0.011760888828171624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,3072,0.008046222229798635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,3584,0.24281599786546496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,6144,0.2148737774954902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,2048,0.029079112741682265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,10240,0.015192889504962497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,51200,1.19686402214898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,4096,0.978881729973687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,512,0.017553778158293832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,256,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,128,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,3072,0.28329510158962673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,51200,0.07805689175923665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,10240,0.07181688812043932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,8192,0.027658666173617046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,16384,0.03289599882231818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,512,0.05058755477269491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,4096,0.03163911236657037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,1536,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,512,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,256,0.007879111501905654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,128,0.017270222306251526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,768,0.030906667311986286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,8192,0.11247111691368951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,12288,0.29677777820163304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,65536,3.207032097710503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,64,0.16164355807834202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,768,0.00811644477976693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,768,0.012439111040698158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,3584,0.07847022347980075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,16384,0.019494222270117868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,3584,0.07184711429807875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,5120,0.02032444377740224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,5120,0.13134400049845377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,3584,0.01993777851263682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,2560,0.007626666790909237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,65536,4.630066765679254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,4096,7.977886623806423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,6144,0.053267555104361645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,64,0.004956444518433677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,1536,0.03086044390996297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,16384,0.014956444501876831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,256,0.01088711122671763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,7168,0.016937777400016785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,128,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,6144,0.07791644334793091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,1024,0.009829333258999718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,1536,0.02188355558448368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,3072,0.29873421457078725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,16384,0.014736000034544202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,1536,0.3240933418273926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,1536,0.006423111177153057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,8192,1.8373493618435328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,768,0.006811555474996567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,5120,0.0761902199851142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,4096,0.1383875608444214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,768,0.005852444304360284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,3072,0.1274684402677748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,51200,0.16470311747656927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,768,0.01315199997689989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,2048,0.03836799992455377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,12288,0.5591324700249566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,2048,0.03338755501641168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,256,0.003105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,1536,0.03188000122706095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,51200,0.22922844356960723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,1024,0.010000000397364298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,3584,0.09416888819800483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,6144,0.10395999749501546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,32,0.002573333266708586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,51200,0.9320835537380643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,6144,0.021558221843507554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,1536,0.009715555442704095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,1536,0.013907555076811047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,64,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,2560,0.007920000288221572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,64,0.002817777709828483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,2048,0.40024622281392414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,256,0.029791110091739233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,65536,0.4050942262013753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,768,0.009752888646390703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,6144,0.01852799952030182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,32,0.03011022342575921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,8192,0.013517333401574029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,512,0.014671999547216626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,8192,0.026543999711672466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,51200,0.8670239978366427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,3072,0.007359111474619971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,32,0.00795022232664956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,768,0.010876444478829702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,12288,0.05835821893480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,512,0.008287111090289222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,32,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,51200,0.06298666530185275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,6144,0.0063866666621632045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,3072,0.083423998620775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,64,0.03826399975352817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,51200,0.0596382220586141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,8192,0.02164799968401591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,4096,0.011388444238238864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,1024,0.19089599450429282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,65536,0.7166168954637316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,64,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,8192,0.21127022637261283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,8192,0.03595644566747878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,32,0.005036444299750858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,1536,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,3072,0.012697777814335294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,2560,0.014175110393100314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,256,0.01421333352724711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,1536,0.041857779026031494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,65536,0.09409422344631618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,10240,0.3639209005567763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,512,0.005432888865470886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,128,0.0028560000161329904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,8192,0.059135112497541636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,3584,0.11672888861762153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,2560,0.00795111142926746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,512,0.057088891665140785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,12288,0.08653333452012803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,512,0.10820800065994263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,10240,0.5101742214626735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,2048,0.10207555691401164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,10240,0.007983110845088959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,10240,0.15668355094061956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,16384,0.5034871101379395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,16384,0.2877359920077854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,1024,0.02033066584004296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,51200,0.2152551147672865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,128,0.00647555540005366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,1536,0.01484266585773892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,512,0.007824000385072496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,4096,0.023295111126369897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,10240,0.01440000037352244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,65536,2.139107598198785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,1024,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,128,0.0070426662762959795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,3584,0.008540444076061249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,2048,0.08293422063191731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,512,0.04070577687687344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,128,0.022118222382333543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,16384,0.2125626669989692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,8192,0.2659084531995985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,2048,0.05645777781804403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,4096,0.38198577033148867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,2048,0.02454311152299245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,512,0.026855111122131348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,10240,0.03633155425389608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,1536,0.06673689021004571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,128,0.005626666463083691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,2048,0.07915021975835164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,256,0.005252444495757421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,7168,0.3125386767917209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,51200,1.0213146209716797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,6144,0.027049778236283198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,32,0.00942311104800966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,512,0.005912000106440649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,65536,0.2551395628187391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,128,0.010877333581447601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,2560,0.06687288814120822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,128,0.22994399070739746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,10240,0.05618488788604736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,3072,0.176292445924547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,128,0.0029493332323100832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,16384,0.13377422756618923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,6144,1.2660889095730252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,2560,0.8978488710191516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,8192,0.01588000026014116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,7168,0.02401422295305464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,2560,0.008811555802822113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,3072,0.29605600568983287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,2048,0.01513155632548862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,6144,0.06277422110239665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,1536,0.005392000079154968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,6144,0.08347022533416748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,8192,0.01681066718366411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,65536,4.164006974962023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,2560,0.8898070653279623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,128,0.005091555416584015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,256,0.012973333398501078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,12288,0.026724444495307073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,1024,0.0262106673585044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,51200,0.7201920085483127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,8192,0.015885333220163982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,7168,0.04830400149027506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,512,0.003964444415436851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,2560,0.00924533357222875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,1536,0.08598311079872979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,1536,0.096596446302202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,1536,4.503504011366102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,256,0.005120888766315249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,128,0.008841777841250101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,3072,0.08246133062574598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,8192,0.3724479940202501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,10240,0.653623104095459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,128,0.006294222341643439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,128,0.003060444361633725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,5120,0.06537777847713895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,2560,0.1290071143044366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,6144,0.017980444762441847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,768,0.0037582222786214617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,64,0.002576888849337896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,768,0.003920888735188378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,6144,0.02437955637772878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,64,0.010928000013033548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,128,0.009519110951158736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,2560,0.013225778109497495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,51200,0.28158754772610134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,32,0.0025724443710512584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,512,0.00462755560874939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,3072,0.2919626765780979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,512,0.0037360000941488477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,4096,0.021673777037196692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,3584,0.013292444248994192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,16384,0.09684977928797404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,12288,0.024728000164031982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,4096,0.1546079980002509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,64,0.03306488858328925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,7168,0.014254222313563028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,16384,0.025250666671329077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,32,0.00556355549229516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,3584,0.1164240042368571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,7168,0.02717511190308465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,128,0.0041680000722408295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,256,0.007595555649863348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,512,0.009465778039561378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,3584,0.006744888921578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,2048,0.0328106681505839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,12288,0.03539377782079909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,3584,0.059883554776509605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,3584,0.5210017628139919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,1536,0.031882666879230075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,10240,0.006028444402747684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,2560,0.018231110440360177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,4096,0.16061155001322427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,2048,0.0178195552693473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,512,0.021665778425004747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,5120,0.2541582319471571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,3584,0.013078221844302284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,1024,0.004880889008442561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,12288,0.07765511009428236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,1536,0.6466959847344292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,128,0.04048177931043837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,4096,0.08601155545976426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,2048,0.04175288809670342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,8192,0.5933849016825358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,2560,0.007792889244026608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,512,0.017991110682487488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,2048,0.030172444052166406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,2048,0.25407645437452525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,8192,0.0536480016178555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,3584,0.01204622205760744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,6144,0.045576890309651695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,768,0.28175200356377494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,5120,1.0286711586846244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,1024,0.00739822205570009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,512,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,1536,0.008271111382378472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,1536,0.3955262237124973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,8192,0.06119644641876221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,32,0.01713511182202233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,12288,0.08268266916275024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,7168,0.11448889308505589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,16384,4.864854176839192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,256,0.010030222435792288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,2048,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,16384,0.7904604276021322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,512,0.004293333325121137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,512,0.020803555846214294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,10240,0.8555786874559191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,128,0.005995555470387141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,3072,0.008924444516499838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,51200,0.1602675517400106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,3584,0.017969777186711628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,12288,0.057633777459462486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,3584,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,256,0.03977955712212457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,4096,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,3072,0.03002133303218418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,2048,0.39382044474283856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,128,0.008591111335489485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,16384,0.7933475706312391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,512,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,32,0.019398222366968792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,2048,0.014366222752465142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,6144,0.014848000473446317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,128,0.008979555633332994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,2048,0.01752800080511305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,4096,0.03619911273320516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,512,0.040864000717798867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,3584,0.03825066818131341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,5120,0.03286844491958618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,1024,0.005393777870469623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,5120,0.08073866367340088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,16384,0.09015466769536336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,7168,1.1774080064561632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,12288,0.6771946483188205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,768,0.27137690120273167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,64,0.003535110917356279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,128,0.027377777629428442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,6144,0.4245439900292291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,12288,0.01793066660563151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,12288,0.029280887709723577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,512,0.08628622028562759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,3072,0.0224906669722663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,8192,0.20119822025299072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,128,0.005480888817045424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,65536,0.05482755435837639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,10240,0.013935999737845527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,8192,0.42682843738132054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,128,0.014327999618318347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,1536,0.005208888815508948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,3584,0.01500266624821557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,512,0.007251555720965068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,768,0.1818924480014377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,256,0.00453333349691497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,3072,0.07362844546635945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,3072,0.006575111299753189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,768,0.005498666730191972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,32,0.005404444618357553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,256,0.14483022689819336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,256,0.005395555661784277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,2048,0.04500266578462389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,1024,0.6398915714687771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,2048,0.016823111308945548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,256,0.0266302227973938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,2048,0.08363288640975952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,64,0.014597333139843412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,256,0.04400177796681722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,7168,0.01255644443962309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,512,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,5120,0.016843555702103507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,12288,0.5716560151841905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,1024,0.005015111217896144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,5120,0.12345955106947158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,4096,0.05925866630342272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,7168,0.0542755557431115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,51200,0.1468071142832438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,256,0.004433777597215441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,16384,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,12288,0.009961778091059791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,8192,0.018811555372344125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,256,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,2560,0.02713955607679155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,32,0.0031555555760860443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,32,0.12787644068400064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,4096,0.04208088914553324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,6144,0.050119111935297646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,6144,0.019854222734769184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,5120,0.05630489190419515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,3584,0.032652444309658475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,10240,0.2187662257088555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,10240,0.3395262294345432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,3072,0.007883555359310573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,1024,0.7213564448886447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,65536,0.36653865708245176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,512,0.005445333404673471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,51200,0.20404088497161865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,16384,0.315595547358195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,4096,0.1760079993142022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,51200,2.819504843817817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,256,0.01882577770286136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,16384,0.16712621847788492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,768,0.004150222159094281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,4096,0.03221599923239814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,3584,0.025848888688617285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,5120,0.10708355241351658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,2560,0.00600000015563435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,16384,0.15335822105407715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,512,0.0058711109062035876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,2048,0.03532800078392029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,1536,0.008818666968080733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,32,0.16021955013275146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,16384,0.5001555548773872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,4096,0.08611377742555405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,256,0.004297777596447202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,51200,0.10327555735905965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,256,0.0033128888656695685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,3072,0.020089776979552377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,64,0.0025840000145965153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,5120,0.026000888811217412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,3072,0.006409777949253718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,51200,42.88824123806424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,12288,3.0347671508789062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,2048,0.3455928961435954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,10240,0.626148435804579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,1024,0.07120711273617215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,2560,0.014158222410413953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,512,0.00886311133702596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,128,0.0047066667013698155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,128,0.010311110979980892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,64,0.025803556044896443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,6144,0.09399111403359307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,768,0.005251555393139522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,64,0.006492444624503453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,5120,0.023711999257405598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,1024,0.012337778177526263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,51200,1.2984417809380424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,16384,0.03952088952064514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,4096,0.011476444701353708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,65536,3.987201690673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,768,0.06682399908701579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,65536,0.2936115529802111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,3584,0.014155555102560254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,5120,0.07421866390440199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,3584,0.006544000158707301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,2048,0.011349333657158745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,6144,0.14963377846611872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,51200,0.16567999786800808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,256,0.0028906667398081887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,10240,0.040786667002571955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,12288,0.18353599972195098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,12288,0.2338319884406196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,128,0.00812000036239624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,2560,0.08967110845777725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,7168,0.030376000536812678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,12288,0.12197777960035537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,5120,0.4594133165147569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,64,0.002630222174856398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,6144,0.016024889217482675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,16384,0.011330666641394297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,8192,2.9139626820882163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,32,0.007312888900438945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,12288,0.23382221327887642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,3072,0.014905777242448596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,512,0.005997333261701796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,7168,0.031011554929945204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,10240,0.020489777127901714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,6144,0.026557332939571802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,768,0.02942044536272685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,512,0.0395066671901279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,7168,0.025880888104438782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,16384,0.03943822118971083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,1536,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,2560,0.04225066635343763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,65536,0.0682577755716112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,512,0.041499555110931396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,12288,0.14785155985090467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,7168,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,1536,0.0046773333516385825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,51200,4.899591233995226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,65536,2.8216364118787975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,128,0.0070977773931291364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,7168,0.33832089106241864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,7168,0.8407582177056206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,128,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,51200,0.06882222493489583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,7168,0.06085155407587687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,5120,0.05617244376076592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,51200,0.07015466690063477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,128,0.00999022192425198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,512,0.0034977777136696708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,65536,1.105603535970052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,51200,0.10304800007078384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,3072,0.009212444225947062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,10240,0.023744000328911677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,128,0.00998044427898195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,12288,0.11921421686808269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,1024,0.15785955058203802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,4096,0.018429333964983623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,6144,0.07973511351479425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,128,0.009328888522254096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,128,0.0203804439968533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,768,0.011041777829329172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,3072,0.016717332932684157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,768,0.012463111016485425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,3584,0.007121777368916406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,64,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,5120,0.015816888875431485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,1024,0.004839999808205498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,2048,0.025046222739749487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,10240,0.013771555489963956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,16384,0.2970355616675483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,6144,0.7811733351813422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,512,0.00904177791542477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,8192,0.1669600009918213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,64,0.027408889598316614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,1536,0.00675644435816341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,2048,0.023557333482636347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,3584,0.057998220125834145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,12288,0.15273243851131862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,5120,0.2977084318796794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,768,0.05214222272237142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,128,0.004471999903519948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,64,0.008547555241319869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,8192,0.014031110538376702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,768,0.003680888977315691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,768,0.00370399985048506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,3072,0.013624889155228933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,51200,0.1425280041164822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,6144,0.08954488568835789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,64,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,51200,4.899242824978299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,2048,0.017458667357762653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,10240,0.14478666252560085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,2560,0.017855111095640395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,32,0.014164444473054675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,768,0.0052284445199701525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,1024,0.026183999247021143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,8192,0.14528711636861166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,3072,0.05810311105516222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,1536,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,128,1.4643964767456055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,8192,0.05786844756868151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,64,0.0026346666531430352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,6144,0.013814222481515674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,512,0.01902844508488973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,7168,0.054416000843048096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,16384,0.1033875544865926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,3584,0.022551111049122278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,65536,1.1407297982109916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,16384,0.037131554550594754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,2048,0.026310222016440496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,6144,0.050943109724256724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,256,0.0030782222747802734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,256,0.0029493332323100832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,51200,0.42247822549608016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,768,1.3593342039320204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,256,0.004359999878538979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,12288,0.022991999983787537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,8192,0.0325617790222168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,512,0.0125946667459276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,128,0.02165422174665663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,8192,0.3029208978017171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,10240,0.030235555436876085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,1536,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,10240,0.029709332519107397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,65536,1.4980986913045247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,12288,0.31948889626397026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,6144,0.05583733320236206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,65536,0.3421822124057346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,7168,0.011659555964999728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,2048,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,32,0.0052897776994440295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,2048,0.027248000105222065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,8192,0.07651733027564155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,3584,0.01888799998495314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,7168,0.014364444547229342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,128,0.0035084444615576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,65536,1.5192168553670247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,4096,0.0361422234111362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,3072,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,1536,0.010462221999963125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,1536,0.026914666096369427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,10240,3.545286178588867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,3072,0.05527111225657993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,6144,0.34551111857096356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,1024,0.04277599851290385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,12288,0.03297422329584757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,2048,0.012879999975363413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,768,0.006688888702127669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,1024,0.01443555619981554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,65536,0.9103928671942817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,1536,0.005236444373925527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,3072,0.009169778062237633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,256,0.012755555411179861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,2560,0.008492444124486711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,512,0.007127111156781514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,2560,0.01312622262371911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,1536,0.15143556065029568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,65536,0.07465866539213392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,768,0.011140444212489657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,1536,0.086081776354048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,65536,0.1921688848071628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,512,0.008532444636027018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,1024,0.019018666611777413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,2560,0.03728355632887946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,768,0.00590844452381134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,65536,0.21700444486406115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,2048,0.10298044151730007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,32,0.005564444594913059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,12288,0.019256888164414298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,10240,0.19960444503360322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,12288,0.12489955955081516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,10240,0.010847110715177325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,2560,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,16384,0.01856355534659492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,51200,0.060645335250430636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,128,0.009083555804358589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,256,0.03946933481428359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,768,0.0125591109196345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,7168,0.07284088929494222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,64,0.14106310738457575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,32,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,4096,0.021977777282396953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,256,0.0029013332807355454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,1536,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,3584,0.0169813334941864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,256,0.013697778185208639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,32,0.006562666760550604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,10240,0.025661332739724055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,65536,0.13581066661410862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,10240,0.049570666419135205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,64,0.00254844439526399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,768,0.029086222251256306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,64,0.00777422222826216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,512,0.020592000749376085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,51200,5.095369974772136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,2048,0.013916444447305469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,2048,0.06884355677498712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,12288,0.019919999771647982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,512,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,51200,3.392322752210829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,65536,0.26747910181681317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,1024,0.2082151174545288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,4096,0.13442844814724392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,16384,0.007921777665615082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,64,0.0034808889031410217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,64,0.002570666579736604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,65536,0.15170578161875406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,1536,0.005028444445795483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,1024,0.04126311010784573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,128,0.004002666721741359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,2048,0.04235288831922743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,7168,0.6640382342868382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,16384,1.5247270796034071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,256,0.016759110821617972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,2048,1.2694471147325304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,64,0.00982844415638182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,5120,0.015190222197108798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,3072,0.013853333062595792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,5120,1.1682666142781575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,12288,0.5399911138746474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,64,0.018497778309716117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,3584,0.057145780987209745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,64,0.014064000712500678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,5120,0.041950222518708974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,3072,0.008656889200210571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,256,0.0058853332367208265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,6144,1.1672301822238498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,16384,0.08027288648817274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,1024,0.004431999805900785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,1536,0.026576000783178542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,128,0.010298666854699453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,2048,0.014560000763999091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,2560,0.014373333917723762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,10240,0.0645244452688429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,16384,0.034887999296188354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,1024,0.05442666676309374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,2560,0.09404799673292373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,6144,0.1343911091486613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,5120,0.03637866510285272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,512,0.03160088923242357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,2560,0.09754488865534465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,5120,0.25289776590135365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,10240,0.024969778127140466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,4096,0.00833511104186376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,2048,0.045124444696638316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,10240,0.048863109615114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,3584,0.013887999786270989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,2560,0.04040266738997565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,128,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,4096,0.016781333420011733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,16384,1.808749304877387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,6144,1.7068381839328344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,4096,0.32299555672539604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,768,0.0041697778635554844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,65536,0.19456266032324898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,5120,0.15853956010606554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,1024,0.02402755618095398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,10240,0.17945688300662568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,51200,5.680499606662327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,12288,0.011419555379284753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,512,0.035154667165544294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,8192,0.0174026671383116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,5120,0.9226480060153537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,2560,0.01089866707722346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,768,0.05915288792716133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,6144,0.014642667439248828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,4096,0.0118568887313207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,5120,0.009824000298976898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,5120,0.07742310894860162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,7168,0.7781626913282605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,10240,0.31284978654649526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,512,0.01981244484583537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,5120,0.11103644635942246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,768,0.00499199993080563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,12288,0.16970045036739775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,64,0.0038453332251972626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,64,0.0034942221310403612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,1536,0.07495555612776014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,6144,0.06468088759316339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,1024,0.052863111098607383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,64,0.028582221931881372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,2560,0.010132444401582083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,51200,2.247274610731337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,65536,6.532928042941624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,7168,0.014000889327791003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,256,0.003699555579158995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,65536,1.1788213517930772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,1024,0.00812977800766627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,768,0.03605955508020189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,65536,3.2584283616807728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,7168,0.018428444862365723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,256,0.0038862224254343244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,8192,0.14368444018893772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,5120,0.14369155300988093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,12288,0.07040266858206855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,7168,0.028336889213985864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,128,0.00452888881166776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,65536,0.5614426930745443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,4096,0.010301333334710863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,65536,0.6639182302686903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,1024,0.020093333390023973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,1024,0.006113777971929974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,7168,0.31788799497816295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,12288,0.5467706786261665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,1536,0.020548444655206468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,512,0.012541333006487953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,12288,0.019652444455358718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,3072,0.011067555182509951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,3584,0.15316800276438394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,10240,0.027473777532577515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,10240,0.06770666440327962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,10240,0.031601776679356895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,256,0.13526311185624865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,1536,0.03105688757366604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,5120,0.13497689035203722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,65536,1.7595831553141277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,1024,0.011507555842399597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,51200,1.1259644826253254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,64,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,768,0.006955555743641323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,1024,0.042056000894970365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,8192,0.02083288960986667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,128,0.004607999904288186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,1536,0.031515555249320135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,4096,0.020586666133668687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,64,0.013271999855836233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,512,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,256,0.012152888708644442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,3584,0.0192239996459749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,128,0.17206044991811117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,128,0.14339733123779297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,32,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,256,0.015638222297032673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,2560,0.042577776643964976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,8192,0.013519110778967539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,2560,0.012897777888509961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,3584,0.6057937939961752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,6144,0.011048888994587792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,1024,0.010850666297806634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,16384,0.533748467763265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,16384,0.03421155611673991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,16384,0.030191110240088567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,5120,0.008103999826643202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,65536,0.1152871052424113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,1536,0.022868444522221882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,512,0.0034382223255104492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,10240,0.006324444380071428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,3072,0.012324444121784635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,6144,0.06990755266613431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,12288,0.10814933644400702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,6144,0.09920089112387763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,64,0.0026835555003749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,768,0.032828443580203585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,128,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,512,0.02942044536272685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,10240,0.03613600134849548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,1024,0.0773457752333747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,8192,0.02641155487961239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,64,0.008231999973456064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,12288,0.017824000782436795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,3072,0.060216890441046826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,256,0.013009777499569787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,128,0.0030702222138643265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,3072,0.009659555223253038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,32,0.0031795555518733132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,16384,0.03489333391189575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,5120,0.0922808912065294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,256,0.005072000126043956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,3584,0.036758220858044095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,64,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,128,0.019671110643280875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,768,0.04020177655749851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,16384,0.023243554764323767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,65536,0.8122488657633463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,2048,0.03904089000489976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,12288,0.007800000409285228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,7168,0.08712266551123725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,16384,0.03142222099834018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,51200,2.962384965684679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,16384,0.7396222220526801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,2560,0.010224888722101847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,3072,0.025994666748576697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,10240,0.038980444272359215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,65536,0.018782221608691745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,8192,0.33941955036587185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,512,0.022638221581776936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,65536,0.909430185953776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,8192,0.01262399968173769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,512,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,3584,0.039044443104002215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,128,0.004549333204825719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,6144,0.18063288264804414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,7168,0.4141955640580919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,4096,0.010427555276287926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,6144,0.03627644313706292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,2560,0.0713955561319987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,65536,0.07513955566618177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,1024,0.022727999422285292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,8192,0.14700622028774685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,16384,0.030435555511050757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,6144,0.19945689042409262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,65536,0.17617332935333252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,512,0.029807110627492268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,512,0.11167644129859076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,7168,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,65536,4.6574889289008246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,8192,0.6442746586269802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,3072,0.08186399936676025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,65536,0.015785778562227886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,51200,0.6023786862691244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,256,0.004514666481150521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,512,0.027273777458402846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,65536,1.886959923638238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,2048,0.008616000413894653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,32,0.003907555507289039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,32,0.0057546666098965546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,3584,0.09181066354115804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,1536,0.006320888797442119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,12288,0.019353777170181274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,7168,0.029231998655531142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,16384,0.12304533852471246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,4096,0.026698667142126296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,16384,0.03570666578080919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,3584,0.024008888337347243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,1024,0.011994666523403592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,8192,0.016136889656384785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,7168,2.0227538214789496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,3584,0.7968053287929959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,32,0.004271111140648524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,7168,0.007387555307812161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,12288,0.6749448776245117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,16384,0.07837777667575412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,65536,0.1800168885125054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,4096,0.09101688861846924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,1536,0.05738933218850029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,51200,0.9950479931301541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,1536,0.07041422526041667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,128,0.004821333206362194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,3584,0.03484977616204156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,2560,0.052632000711229115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,512,0.17617066701253256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,2560,0.10213600264655219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,65536,2.569729699028863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,5120,0.10910133520762126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,64,0.0046924443708525766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,3584,0.01625155574745602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,12288,0.026662222213215295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,3584,0.01203644441233741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,2560,0.00846400029129452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,10240,0.39407555262247723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,7168,0.025050666597154405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,3072,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,7168,0.2853599919213189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,65536,0.36816800965203184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,6144,0.020489777127901714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,6144,0.7512577904595269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,3584,0.007838221887747446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,128,0.0027591110103660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,51200,0.05722755855984158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,64,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,5120,0.058783113956451416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,512,0.004352000024583605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,5120,0.05528710948096382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,5120,0.12502045101589626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,12288,0.03016977839999729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,10240,0.045787556303872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,512,0.04351910948753357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,768,0.005444444302055571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,2048,0.007540444533030192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,256,0.004883555488453971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,128,0.005149333427349727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,512,0.0037431112594074676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,10240,0.041907555527157254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,768,0.043579555220074125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,1536,0.0069102222720781965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,3584,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,3072,0.010168888502650792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,2560,0.09961777925491333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,10240,0.28442133797539604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,5120,0.1410035557217068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,16384,0.017130666308932833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,2048,0.008574222524960836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,16384,0.05298666821585762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,256,0.0029475556479560006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,3584,1.4784764183892145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,2560,0.012952889005343119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,3072,0.010079999764760336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,8192,0.07361244493060641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,65536,0.07530311081144545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,128,0.0046640001237392426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,256,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,1024,0.006581333362393909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,64,0.006709333509206772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,512,0.0032622222271230486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,4096,0.017519111434618633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,4096,1.5527110629611547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,32,0.13313955730862087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,10240,0.023510222633679707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,6144,0.01568533314598931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,51200,4.730772654215495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,12288,0.04266311062706841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,3072,0.07724977864159478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,5120,0.12916355662875706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,2560,0.14249332745869955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,4096,0.008933333059151968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,7168,0.06042666567696465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,128,0.006083555519580841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,51200,0.1395182212193807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,1536,0.004559111264016894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,65536,0.5304222106933594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,8192,0.6072986920674642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,4096,0.027855998939938013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,7168,0.015810666812790763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,1536,0.026524444421132404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,10240,0.9330381817287869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,768,0.020222221811612446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,2560,0.007382222347789341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,768,0.00871288859181934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,7168,0.9248853259616427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,7168,0.022309333086013794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,4096,0.01145244472556644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,512,0.010082667072614035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,3072,0.010891555911964841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,3072,0.007466666400432587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,32,0.005124444348944558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,10240,0.015473778049151102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,1024,0.044383999374177724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,2560,0.14192267258961996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,7168,1.2617884741889107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,3072,0.012508444488048553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,5120,0.0986115535100301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,512,0.005346666607591841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,64,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,1024,0.047247111797332764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,65536,6.512322743733724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,2560,0.11496088239881729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,256,0.013905778527259827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,1024,0.033424887392256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,8192,0.020063999626371596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,3072,0.007128000259399414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,1536,0.015631111131774057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,256,0.0221377776728736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,4096,0.05198755529191759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,2560,0.017495999733606975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,12288,0.01848266687658098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,1536,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,4096,0.019253333409627277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,64,0.011539555258221097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,51200,0.19291467136806914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,51200,0.5010648833380805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,2048,0.017889777819315594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,64,0.005024000174469418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,1024,0.005767999837795894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,512,0.003096888876623578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,8192,0.3802186648050944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,16384,0.4962204297383626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,10240,0.22419910960727266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,16384,0.03979288869433933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,768,0.023846222294701472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,512,0.005436444448100195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,65536,0.8087324566311307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,2560,0.045737776491377086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,128,0.007506666911972894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,5120,0.08444533083173965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,1024,0.011142222417725457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,4096,0.0917137795024448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,6144,0.01532977819442749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,3072,0.04749511016739739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,7168,0.013188444077968597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,12288,0.052207999759250216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,768,0.012472888661755456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,64,0.005010666532648934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,16384,0.022321777211295232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,768,0.03695377707481384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,51200,0.4135564433203803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,2560,1.5504364437527125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,12288,0.02438933319515652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,12288,0.32416799333360463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,1536,0.007074666519959767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,1024,0.010286221901575724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,2560,0.008008889026112026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,3072,0.010055111514197456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,51200,1.4959146711561415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,8192,3.208041720920139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,2048,0.04208088914553324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,128,0.0038008888562520346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,32,0.0028000000036425064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,51200,0.4239777723948161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,4096,0.06870844629075792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,512,0.003137777869900068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,256,0.0053395554423332214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,6144,0.1012071106168959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,1536,0.021516443954573736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,32,0.017286222842004564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,768,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,12288,0.05745689074198405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,64,0.011214222345087262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,512,0.004359111189842224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,12288,0.027622222900390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,2048,0.009319110876984065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,12288,0.48507732815212673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,7168,0.3698177867465549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,6144,0.011242667006121742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,1024,0.045033779409196645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,51200,0.01475911173555586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,65536,0.7758657667371961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,1536,0.025787555509143408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,5120,0.0476453337404463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,256,0.020749333831999037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,3072,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,5120,0.014160000615649752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,7168,0.05781955851448906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,128,0.05333422289954292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,7168,0.38282577196757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,12288,0.030226667722066242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,65536,0.1229048834906684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,2048,0.008631111019187504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,1024,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,64,0.010779555473062726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,3072,0.009621333744790819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,32,0.02204622162712945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,128,0.002895999906791581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,3584,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,4096,0.018704000446531508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,51200,0.10266577535205418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,2560,0.011176000038782755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,65536,0.7583955658806695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,256,0.021753778060277302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,65536,4.920587751600477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,1536,0.03197600113021003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,32,0.006464889066086874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,65536,0.7532613542344836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,10240,0.04072177741262648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,3072,0.052816000249650746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,8192,0.4452968703375922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,2048,0.15765778223673502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,1024,0.007246221933099959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,8192,0.06270311276117961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,768,0.2506311204698351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,4096,0.022314666046036616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,512,0.004432888908518685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,7168,0.026692445079485577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,128,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,2560,0.0064106666379504735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,512,0.005736888696750005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,128,0.003014222201373842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,16384,0.046167999505996704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,65536,2.8888195885552297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,512,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,5120,0.012903110848532783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,4096,0.2068666617075602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,6144,0.023941333095232647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,2048,0.0068604445291890045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,65536,0.0403386652469635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,1536,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,4096,0.022223111655977037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,2560,0.010467555787828235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,12288,0.08698399861653645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,51200,5.493159823947483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,256,0.010495999621020423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,4096,0.05985689163208008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,12288,5.376719156901042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,768,0.10145333078172471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,256,0.022724444667498272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,1536,0.006462222172154321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,3072,0.007609777980380588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,3072,0.04467466804716322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,10240,0.047190222475263804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,7168,0.02983377708329095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,512,0.009952888720565373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,51200,0.17187733120388457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,3584,0.01958488921324412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,12288,0.023157333334287006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,4096,0.07598755757013957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,5120,0.7300444708930122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,768,0.05137333273887634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,3072,0.22161600324842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,3072,0.15054399437374538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,2560,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,5120,0.07460177607006498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,7168,0.8377075725131564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,128,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,4096,0.06262933545642428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,8192,0.10572089089287652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,12288,0.03143644332885742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,6144,0.035862220658196345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,2048,0.3699928919474284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,51200,0.031168000565634832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,4096,0.03280355532964071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,32,0.003559111307064692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,64,0.1795253356297811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,3072,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,12288,0.21859822008344862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,768,0.005953777581453323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,8192,0.03488622109095255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,10240,0.08609066406885783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,4096,0.020276443825827706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,1024,0.017941332525677152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,5120,0.015485333071814643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,12288,0.5766008694966634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,6144,0.3863182332780626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,3584,0.6226302252875434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,3072,0.00959644466638565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,4096,0.009539555344316695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,128,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,10240,0.027461333407296076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,1536,0.01864355636967553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,8192,7.051076253255208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,2560,0.05658666955100166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,4096,0.05529510974884033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,3072,0.026296888788541157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,12288,0.03046577837732103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,3072,1.2756995095147026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,7168,0.6101919809977213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,7168,0.06804177496168348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,5120,0.009514666265911525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,7168,0.012106666962305704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,16384,0.07032710976070829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,1536,0.025806221697065566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,768,0.08774399757385254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,1024,0.009032888544930352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,1024,0.006626666833957036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,1536,0.00831200016869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,32,0.005342222336265776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,2560,0.007632888853549957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,768,0.025800888737042744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,7168,0.012883555557992725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,3072,0.019222221440739103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,1024,0.015135111080275642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,4096,1.1199377907647026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,7168,0.21316799852583143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,12288,0.12403289477030437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,128,0.0029848888516426086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,128,0.005351999981535806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,7168,0.009641778137948778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,256,0.005577777822812398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,65536,0.26989245414733887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,65536,0.06978933016459148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,8192,0.017470222380426195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,51200,0.0661768913269043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,2560,0.009020444419648912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,3072,0.010564444793595208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,256,0.01457333399189843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,3584,0.01795644395881229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,3584,0.05659911367628309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,12288,0.8176088862948947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,5120,0.3959848880767822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,10240,0.0395760006374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,3072,0.23267110188802084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,2048,0.027965333726671007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,3584,0.3827982213762071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,256,0.006113777971929974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,768,0.01035199976629681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,512,0.006903999795516332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,5120,0.32802221510145396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,128,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,8192,1.2616088655259874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,12288,1.4834008746676977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,128,0.005048000150256687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,51200,0.17453155252668592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,3072,0.009986666341622671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,51200,15.838587443033854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,64,0.031346665488349065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,10240,0.015610666738616096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,65536,0.07010133398903741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,1024,0.035713778601752386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,4096,1.0278426276312935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,7168,0.014249778456158109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,64,0.015440889530711703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,65536,0.1286968919965956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,128,0.0047173334492577445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,10240,0.018556444181336295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,4096,0.04433777928352356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,64,0.004050666673315896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,6144,0.01903555625014835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,16384,0.07327822181913587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,3072,0.050125333997938365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,256,0.02581688927279578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,2560,0.013940445250935025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,2560,0.007596444752481248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,5120,0.02793688906563653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,65536,1.855002721150716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,1536,0.13211377461751303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,512,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,3072,0.006689777804745569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,3072,0.009367110828558603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,256,0.003185777821474605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,7168,0.0742480026351081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,6144,0.05990933047400581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,4096,0.053653332922193736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,3584,0.008531555533409119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,8192,0.044863111442989774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,1024,0.043078223864237465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,64,0.0053084443012873335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,2560,0.06935022274653117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,512,0.005623111294375525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,16384,0.014608888162506951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,5120,0.09675111373265584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,65536,0.1314444409476386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,64,0.006496888895829518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,7168,0.10473333464728461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,10240,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,512,0.006090666684839461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,3072,0.035381333695517644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,10240,0.018008889423476327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,5120,2.668026606241862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,5120,0.35911199781629777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,256,0.014352889524565803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,128,0.009395555489593083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,6144,0.1207324398888482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,1024,0.0107351111041175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,64,0.014452444182501899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,3072,0.008002666963471307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,3584,0.06611822048823039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,65536,0.2931191126505534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,12288,0.017462222112549674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,128,0.00978488889005449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,51200,0.4417102336883545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,12288,0.23407999674479166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,65536,3.2739707099066844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,6144,0.016327111257447135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,12288,0.12294844786326091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,32,0.003032888803217146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,768,0.01556799974706438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,10240,0.15653244654337564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,512,0.00887022250228458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,12288,0.04650400082270304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,16384,0.525968869527181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,8192,0.03161777721510993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,5120,0.012154666913880242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,1536,0.008772444393899705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,512,0.006210666563775804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,256,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,6144,2.1239634619818792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,7168,0.0107360002067354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,1536,0.026151999831199646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,12288,0.026485333840052288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,3072,0.10994933048884074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,2560,0.08428888850741917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,10240,0.19436711735195586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,256,0.0058293334311909145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,2560,0.010460444622569613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,64,0.002793777734041214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,256,0.01776622235774994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,5120,0.07794844441943698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,8192,0.6795502238803439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,1536,0.01940088967482249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,2048,0.03396888905101352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,12288,0.055721779664357506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,6144,1.5991413328382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,768,0.010832000109884473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,1024,0.2979493406083849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,5120,0.10078755352232192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,7168,0.056742224428388804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,65536,1.703993797302246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,512,0.02028622229894002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,16384,0.24725243780348036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,7168,0.01891377733813392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,1536,0.009514666265911525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,128,0.0028311111446883944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,5120,0.0629075566927592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,16384,0.08164266745249431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,7168,0.1210906638039483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,3072,0.10580622487597996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,1024,0.004064889003833135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,7168,0.01792711185084449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,7168,0.27262843979729545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,4096,0.007920888562997183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,2560,0.019095111224386428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,51200,0.45791467030843097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,4096,0.062183108594682485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,128,0.006550222221348021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,256,0.041308446062935725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,5120,0.0514355566766527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,65536,0.4185733265346951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,2048,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,4096,0.013732444081041547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,65536,0.08086311154895358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,1536,0.5474088986714681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,64,0.004886222382386525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,51200,0.14934578206804064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,512,0.008310221963458592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,6144,0.15952000353071424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,6144,0.02981511089536879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,6144,0.2004399961895413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,3072,0.8905671437581381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,2560,0.01564800077014499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,5120,0.07116711139678955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,6144,0.007374222079912822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,8192,0.024799111816618178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,16384,0.17599466111924914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,12288,0.03052088949415419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,2048,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,32,0.013659555878904132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,1536,0.14556977483961317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,256,0.2419439951578776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,65536,0.23622311486138237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,10240,0.04166488846143087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,2048,0.007199111084143321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,2048,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,65536,0.10284444358613755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,2560,0.014205333259370593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,1536,0.02971733278698391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,128,0.004607111215591431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,2048,1.9230435689290364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,4096,0.014448000325096978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,16384,0.09792533185746934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,51200,2.0637075636121964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,6144,0.0473235547542572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,16384,0.024672889047198828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,8192,0.014354666074117025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,1536,0.08287644386291504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,6144,0.038972444004482694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,512,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,16384,0.09425155321756999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,768,0.017299556069903903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,5120,0.030797332525253296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,6144,0.09964711136288112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,16384,0.01832177738348643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,6144,0.007302222152551015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,7168,0.5292133225335015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,3584,0.31999821133083767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,1536,0.032840887705485024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,128,0.030095110336939495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,65536,0.29738577206929523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,8192,0.06721066766315036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,512,0.017320000463061862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,7168,0.0428693327638838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,16384,6.569537692599827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,256,0.12863822778066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,512,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,65536,0.20545600520239937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,1024,0.022273777259720698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,12288,0.02476088868247138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,3584,0.018726223044925265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,3584,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,7168,1.173000865512424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,5120,0.011665778027640449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,2048,0.07967733012305366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,1536,0.010082667072614035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,2048,0.010084444450007545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,32,0.004997333304749595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,16384,0.019098665979173448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,6144,0.012697777814335294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,8192,0.32488621605767143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,6144,0.0120666664507654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,65536,0.08913599782519871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,6144,0.9315733379787869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,5120,0.014704000618722705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,7168,0.040977779361936785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,1536,0.02537422213289473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,5120,0.9931892818874783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,8192,0.625340461730957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,2560,0.010543111297819348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,3584,0.009472000102202097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,768,0.035411556561787925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,8192,1.1549484464857314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,2048,0.03711822297837999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,7168,0.33150309986538357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,65536,0.20415288872188994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,12288,0.0888026687833998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,3072,0.10152888960308498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,512,0.005764444255166584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,7168,0.19683555761973062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,3072,0.3429599867926703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,256,0.004464888738261329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,5120,0.007558222446176741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,256,0.002895999906791581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,64,0.002809777855873108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,256,0.007535999847782983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,3584,0.05874933136834038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,51200,0.07633599970075819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,3584,0.055361777544021606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,1536,0.006485333459244833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,1536,0.006229333165619109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,51200,0.24317600991990831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,512,0.012347555822796293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,1536,0.5286755561828613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,2048,0.0464915566974216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,6144,0.2788168854183621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,7168,0.03141511148876614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,16384,0.8544453514946831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,512,0.004964444372389051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,768,0.01276000009642707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,768,0.007429333196745978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,7168,0.03752622339460585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,3584,0.15831644005245632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,512,0.02093688812520769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,128,0.018177777528762817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,64,0.02940800123744541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,3584,0.008931555681758458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,768,0.011715555356608497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,1536,0.061992890304989286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,6144,0.009745777481132084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,10240,0.1512168910768297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,1024,0.1847511132558187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,3072,0.01997244523631202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,65536,1.0809955596923828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,65536,4.456375969780816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,65536,0.10146933131747776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,2048,0.010421333213647207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,16384,0.5450835757785374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,51200,0.25047021441989475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,2560,0.05619733201132881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,3072,1.1316799587673612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,3072,0.011256888508796692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,256,0.004879111217127906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,16384,0.07633155584335327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,10240,0.011622221933470832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,12288,0.057519111368391246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,3584,0.012281777958075205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,65536,2.0674311319986978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,8192,0.045383999745051064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,2048,0.047375109460618764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,6144,0.02125955621401469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,5120,0.32883911662631565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,7168,0.1039759980307685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,7168,0.06782844331529406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,10240,0.04684799909591675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,1536,0.009269333548016017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,3072,0.006777777853939269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,128,0.19286666976081002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,12288,0.05838133229149712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,1536,0.022799111074871484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,32,0.007420444654093848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,10240,0.04331644376118978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,2560,0.009112888740168678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,3072,0.4113608996073405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,10240,0.45679553349812824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,5120,0.015392000476519266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,128,0.005803555664088991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,5120,0.016813332835833233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,4096,0.018696000178654987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,16384,16.810863918728298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,64,0.08663288752237956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,7168,0.15450933244493273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,1024,0.026696888936890498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,5120,0.008894222478071848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,10240,0.14653955565558538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,256,0.005692444327804778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,1536,0.00927022182279163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,51200,0.11864533689286973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,256,0.0029884444342719186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,16384,0.2563848760392931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,51200,1.3321492936876085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,65536,1.2791190677218967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,32,0.014404444230927361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,16384,0.025742222865422566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,8192,2.7996915181477866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,65536,0.06395288970735338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,12288,0.008773333496517604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,7168,0.01554844445652432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,256,0.004488888714048597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,1024,0.00924444446961085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,2048,0.013076444466908773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,8192,0.015560888581805758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,768,0.03775555557674832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,256,0.0032604444358083936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,65536,0.7446728812323676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,3072,0.08161866664886475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,10240,0.4157591130998399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,65536,0.1063235534562005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,10240,0.02662311163213518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,512,0.014797333214018079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,51200,0.36960532930162215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,3072,0.1570728884802924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,3584,0.019131556153297424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,51200,0.013838222457302941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,7168,0.7094320191277398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,2560,0.11578310860527886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,5120,0.2877039909362793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,10240,0.04182044333881802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,32,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,512,0.0856177806854248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,512,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,10240,0.04038133223851522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,512,0.024481778343518574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,6144,0.014993778533405729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,3584,0.015052444405025907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,51200,0.10702666971418594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,64,0.008032889001899296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,4096,0.7917848692999946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,12288,0.6652862230936686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,3584,0.0163093325164583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,128,0.0064275554484791225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,10240,0.0181048893266254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,3584,0.018176888426144917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,65536,0.2574746608734131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,51200,1.3906097412109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,3584,0.018940443793932598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,4096,0.0821031133333842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,12288,0.01899466746383243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,3584,0.009310222334331935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,10240,0.040744000011020236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,4096,0.06206044223573473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,512,0.009202666580677032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,12288,0.02741866641574436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,512,0.0041288890772395665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,7168,0.0564017759429084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,3072,0.07393066750632392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,51200,0.1519120004442003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,4096,0.06505333052741157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,768,0.005838222387764189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,128,0.004436444491147995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,8192,0.04055555661519369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,6144,0.009727111293209923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,512,0.012192000117566852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,3072,0.0874977774090237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,12288,0.5934302012125651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,3072,0.03975733452373081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,2048,0.008298666940795051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,1024,0.06093600061204699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,8192,0.03609422180387709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,2560,0.007530666887760162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,256,0.023200889428456623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,5120,0.022874666584862605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,32,0.005670222143332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,1536,1.018927150302463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,4096,0.01053066634469562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,768,0.007061333292060428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,51200,26.5238037109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,768,0.03755555550257365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,768,0.0045422220395671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,2560,0.01978755493958791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,64,0.007297777467303806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,1536,0.6502222485012478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,1536,0.012964444855848948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,3072,0.014911999305089315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,256,0.02176355487770504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,3072,0.023783110909991797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,12288,0.01812622282240126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,65536,0.1959573295381334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,8192,0.02197333342499203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,2048,0.030576888057920668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,51200,0.2900071144104004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,512,0.005528888768619961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,32,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,3584,0.04418311185306973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,768,0.0051857779423395796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,1536,0.043783999151653714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,2048,0.18149422274695504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,10240,1.359500461154514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,512,0.004646222210592694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,16384,0.0899146662818061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,16384,0.02424000038041009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,512,0.0038782221575578055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,4096,0.009501333038012186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,5120,0.015014222926563688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,3072,0.010997333460383944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,512,0.006177777631415262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,10240,6.5980784098307295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,128,0.00397866674595409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,1536,0.008995555341243744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,768,0.005913777897755305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,512,0.019691555036438834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,51200,0.10983377695083618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,512,0.010079111489984725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,16384,0.11392177475823297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,256,0.006060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,10240,0.2999511029985216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,6144,0.0734142197502984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,3584,0.010451555252075195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,64,0.005106666849719154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,1024,0.05471555391947428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,256,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,2048,0.0052426668504873914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,512,0.025049777494536504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,2560,0.24245243602328828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,51200,0.07295466793908013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,32,0.0203804439968533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,64,0.02509244448608822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,10240,0.08032888836330838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,8192,0.017097777790493436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,6144,0.013794666363133324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,128,0.006627555522653792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,3584,0.011995555626021491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,128,0.024099555280473497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,256,0.033788445923063494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,2048,0.5284978018866645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,51200,0.1367164452870687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,65536,0.23596178160773384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,64,1.519356409708659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,1536,0.006928000185224745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,12288,0.04340799980693393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,8192,0.0566177765528361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,12288,2.3118854098849826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,1536,0.006787555499209299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,32,0.0028773333049482773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,5120,0.0158942225906584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,768,0.005088888936572605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,128,0.005421333428886201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,32,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,2048,0.005383111122581694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,128,0.004363555461168289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,128,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,2560,0.03625155488650004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,7168,0.026808889375792608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,1536,0.016539555456903245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,51200,0.06350666946834989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,16384,0.04035288757748074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,7168,0.016388444436921012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,1536,0.6474880112542046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,4096,0.042930665943357676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,12288,0.2331626680162218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,64,0.006862222320503659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,5120,0.7445377773708768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,6144,0.32596177524990505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,6144,0.2193759944703844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,7168,0.0766506658660041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,64,0.0027226667023367355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,51200,0.020918221937285528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,1536,0.019337778290112812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,5120,0.00656444455186526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,8192,0.9379768371582031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,10240,0.034889777501424156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,128,0.005072000126043956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,12288,3.954022301567925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,128,0.046873778104782104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,64,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,1024,0.0075786668393347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,12288,0.41900889078776044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,2048,0.013628444737858243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,10240,0.03683733277850681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,2560,2.6545236375596786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,3584,0.008375110725561777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,32,0.0032266666077905228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,3072,0.012243555651770698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,7168,0.03310133351220025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,51200,0.16925866074032256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,65536,0.3300008773803711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,128,0.00453599997692638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,12288,0.03673155440224542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,64,0.00832799987660514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,5120,0.05691822369893392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,1536,0.09539200199974908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,51200,1.5691341824001734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,7168,0.015970667203267414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,4096,0.01940088967482249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,128,0.04166222280926175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,128,0.010104888843165504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,16384,0.40570489565531415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,512,0.008088889221350351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,8192,0.006303999986913469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,1024,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,65536,2.2028791639539933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,8192,1.7270515229966905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,5120,0.20636533366309273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,16384,0.6918915642632378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,3072,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,1536,0.032327112224366926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,3072,1.0040853288438585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,5120,0.01992266707950168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,51200,0.15346044964260525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,256,0.00545155546731419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,4096,0.012617777619096967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,3072,0.015192000402344598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,3072,0.007715555528799693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,1536,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,6144,0.0299582216474745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,256,0.007947555846638149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,2560,0.007231999602582719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,1024,0.005802666561471091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,2560,0.1784364382425944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,8192,0.02260711126857334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,65536,0.31649332576327854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,1536,0.015124445160230001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,2048,0.023036445180575054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,3584,0.011515555282433828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,4096,0.01372977760103014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,4096,0.012520888613329994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,65536,0.10159111022949219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,6144,0.05743910868962606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,64,0.0027635554886526535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,512,0.007556444240940942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,64,0.011408000356621213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,1536,0.00833777752187517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,1024,0.03355733222431607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,512,0.020410666863123577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,32,0.010808889236715106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,1024,0.03661688831117418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,768,0.05036088824272156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,1536,0.006341333190600078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,64,0.007280000381999546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,5120,0.02293333411216736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,8192,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,64,0.007495111061467065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,7168,0.10281599892510308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,1024,0.00738666703303655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,8192,0.024108444650967915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,64,0.003403555601835251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,12288,0.10660977496041192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,512,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,10240,0.06769155793719821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,10240,0.01351466692156262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,1536,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,51200,0.6499742401970757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,128,0.03683022326893277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,7168,0.021026665965716045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,1536,0.009147555463843876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,128,0.0028284444577164124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,128,0.0027840000887711844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,256,0.002879111096262932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,512,0.08339733547634548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,2560,0.06586844391292997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,512,0.00349866681628757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,8192,0.12057688501146104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,8192,0.009716444545321995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,8192,0.12552889188130698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,7168,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,512,0.008391111261314815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,128,0.003144888828198115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,512,0.0031146665828095544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,256,0.004429333325889376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,2560,0.06861688693364461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,6144,0.041209777196248375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,128,0.009688888986905416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,2560,0.007352888584136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,12288,0.08678666750590007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,128,0.009755555954244401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,3072,0.1485475566652086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,51200,0.09911111328336929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,16384,0.338135110007392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,3072,0.012327111429638334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,1536,0.024523556232452393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,3584,2.1191991170247397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,2048,0.022110222114457026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,65536,16.871307373046875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,1024,0.007176888485749562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,3072,0.009554666777451834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,768,0.008330666356616551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,128,0.008921778036488427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,1024,0.0075084442893664045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,3072,0.4137057728237576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,4096,0.05503822366396586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,8192,2.0302622053358292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,1024,0.009602666729026372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,6144,0.16128178437550864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,7168,0.01722577710946401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,12288,0.3845591015285916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,10240,0.015263110399246216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,768,0.011310222248236338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,2048,0.007703111403518253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,6144,0.01538755496342977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,4096,0.137809779908922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,64,0.007537778053018782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,2560,0.01755911111831665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,12288,0.08491555849711101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,32,0.006776000062624614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,12288,0.5199244287278917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,12288,0.03461422191725837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,51200,0.11060088872909546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,10240,0.035659554931852556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,768,0.010091555615266165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,2560,0.013556444810496436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,12288,0.24919022454155815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,1536,0.02753955622514089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,32,0.030778666337331135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,65536,1.4865395228068035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,4096,0.06544888681835599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,128,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,768,0.01183199965291553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,3072,0.1377226644092136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,256,0.032807111740112305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,128,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,2048,0.11233866214752197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,16384,0.019150222341219585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,768,0.008029333419269985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,8192,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,51200,0.09802755382325913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,128,0.005146666533417172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,768,0.012301333248615265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,5120,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,256,0.003478222423129611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,2560,2.101959016587999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,65536,4.7360178629557295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,2560,0.007800000409285228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,2560,0.03597777750757005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,256,0.0031591111587153543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,51200,0.39189423455132383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,128,0.01442577772670322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,65536,1.704893324110243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,51200,12.199772304958769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,3072,0.01686844395266639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,256,0.008375999828179678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,12288,0.3931457731458876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,65536,1.0257849163479276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,3584,0.07164711422390409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,16384,0.025705777936511572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,1024,0.014228444960382251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,1024,0.03576711151334975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,256,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,10240,0.011630222201347351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,10240,0.04028088847796122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,2560,0.011770666473441653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,2048,0.1997742255528768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,4096,0.020084444019529555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,768,0.010601777997281818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,3072,0.07433511151207818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,2048,0.007024889190991719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,6144,0.009949333137936061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,5120,0.015698666373888653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,1536,0.09991021951039632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,16384,0.1290808916091919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,512,0.01612355477280087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,256,0.00301155551440186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,65536,0.012270222107569376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,1024,0.02477955487039354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,256,0.03172711200184292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,64,0.00282577777074443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,7168,0.010842666857772403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,10240,0.1318302286995782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,1536,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,4096,0.01570399933391147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,1536,0.00958577791849772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,1536,0.01625866691271464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,256,0.0047395556337303585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,65536,0.5827280150519477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,1536,0.006504000061088138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,6144,0.24382400512695312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,4096,0.0631128880712721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,2560,0.011494222614500256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,65536,0.07144355773925781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,32,0.012737777498033313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,256,0.004286222159862518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,51200,1.48834048377143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,65536,0.19103377395206025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,256,0.0031235555393828284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,256,0.004134222037262387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,3584,0.011561777856614856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,1536,0.013234666652149625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,10240,0.044256889157825045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,768,0.0037493333220481873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,65536,0.5624106725056967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,10240,0.04000622365209792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,12288,0.02699377801683214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,7168,0.08179199695587158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,3584,0.016897777716318767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,16384,0.03596622082922194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,1024,0.014707555373509726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,1024,0.0063582224150498705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,10240,0.16824266645643446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,768,0.006154666758245892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,12288,0.018698667486508686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,5120,0.043693333864212036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,128,0.12519466876983643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,10240,0.017253332667880587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,512,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,3072,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,16384,0.06270133124457465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,51200,1.662857797410753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,51200,0.0618106656604343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,768,0.017806222041447956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,6144,0.012166221936543783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,768,0.03616177704599168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,2048,0.011219555305110084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,128,0.004105777790149053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,10240,1.0036933686998155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,6144,0.32445422808329266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,4096,0.020047111643685233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,3584,0.10875733031166925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,8192,0.015187554889255099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,4096,0.008783111141787635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,16384,0.06286933024724324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,7168,0.021320000290870667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,1024,0.026779555612140234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,7168,0.01907466683122847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,128,0.003986666599909465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,512,0.1959271166059706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,768,0.027804444233576458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,1536,0.009944000177913243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,3072,0.006676444576846228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,256,0.014717333846622042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,6144,0.41425334082709414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,128,0.0162471118900511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,6144,0.25217334429423016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,16384,0.7295226520962186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,1024,0.0421377784676022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,6144,0.01757422255145179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,768,0.008742222355471717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,5120,0.1356417735417684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,128,0.007717333734035492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,6144,0.0477422210905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,10240,0.027756444282001917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,256,0.023011555274327595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,768,0.005756444401211209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,2560,0.006715555571847492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,512,0.006272000157170826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,1536,0.008187555604510838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,12288,0.02555999987655216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,3072,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,768,0.02818044357829624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,2048,0.016028443972269695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,256,0.06639199786716037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,5120,0.022249778111775715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,51200,0.15575110912322998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,32,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,512,0.003439999909864532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,5120,0.010772444307804108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,2048,0.5808862050374349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,128,0.023669333921538457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,3072,0.01903288894229465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,16384,0.020489777127901714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,128,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,7168,0.5085235701666938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,65536,0.053547554545932345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,2048,0.011491555306646558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,51200,0.427349328994751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,4096,0.019221333993805777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,3072,0.05149244599872165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,768,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,7168,0.07512178023656209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,65536,0.2980853451622857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,8192,0.03695822093221877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,12288,0.02480711042881012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,12288,0.04570133487383524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,51200,0.049139555957582265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,4096,0.009934222532643212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,1536,0.005585777676767773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,512,0.030607110924190942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,65536,0.09114044242435032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,5120,0.058315555254618325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,1024,0.05398933423890007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,51200,0.06150222486919827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,51200,1.31751463148329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,3072,0.015358222855461968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,7168,0.21334666675991484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,16384,0.03103644318050808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,32,0.0053386667536364655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,2048,0.027865777413050335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,51200,0.2289146582285563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,7168,0.04691466689109802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,768,0.005752000129885144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,16384,0.0506551100148095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,3072,0.01068800025516086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,16384,0.23612711164686415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,16384,1.4133520126342773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,10240,0.03760977917247348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,8192,0.610330687628852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,5120,0.011738667057620155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,4096,0.011395555403497485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,16384,0.03772977656788296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,7168,0.046430223517947726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,16384,0.05187911126348707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,16384,0.1251253286997477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,10240,0.008940444224410588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,5120,0.0933235552575853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,64,0.0027235555979940626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,10240,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,51200,1.3490915298461914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,512,0.015384889311260648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,10240,0.3298977745903863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,2048,0.018075555562973022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,1536,0.014148443937301636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,3072,0.015045333239767285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,12288,0.33826756477355957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,2048,0.008347555167145198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,768,0.014053333136770459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,12288,0.22644265492757162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,256,0.004519999855094486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,768,0.006177777631415262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,5120,0.016347555650605094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,3584,0.017557332913080852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,10240,0.7972222434149848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,2560,0.006083555519580841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,7168,0.018340445227093168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,3072,0.010076444182131026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,1536,0.015198222464985318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,5120,0.12940443886650935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,768,0.011977777712874942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,3584,0.04563822348912557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,3072,0.20408711168501112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,64,0.0032675556010670135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,2560,0.007363555332024892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,16384,0.036678221490648054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,8192,0.012680889003806643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,2048,0.10140533579720391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,12288,2.3011343214246964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,5120,0.02793244520823161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,3584,0.018616000811258953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,1024,0.10593955384360419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,2560,0.3461137877570258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,32,0.0029173334025674393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,7168,0.061362663904825844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,51200,0.1502844492594401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,3072,0.012338666452301873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,512,0.0075777777367168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,128,0.004222222086456087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,3072,1.039165284898546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,1536,0.30451999770270455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,768,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,64,0.00544177782204416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,3584,0.016723554995324876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,512,0.006160000132189856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,10240,0.01960444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,1536,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,256,0.0033582221302721235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,32,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,512,0.0037271111375755737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,2048,0.005658666706747479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,5120,0.014398222168286642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,65536,0.09579200214809841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,12288,0.028314666615592107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,3072,0.017437333861986797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,4096,0.008973333570692275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,2560,0.010201777848932479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,10240,0.013822221921549903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,128,0.005165333135260476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,7168,0.012472888661755456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,51200,0.13348889350891113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,7168,0.055198222398757935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,256,0.033687111404207014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,2048,0.07686755392286512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,6144,0.19527910815344918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,16384,0.04015555646684434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,51200,10.191899617513021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,16384,0.0081404447555542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,128,0.0052062223354975385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,7168,0.09303555885950725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,51200,0.012868444124857584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,5120,0.014111111561457316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,1024,0.006511999915043513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,512,0.07837600178188749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,12288,0.10825155840979682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,7168,0.06153955724504259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,3072,0.01594577729701996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,10240,0.015538665983412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,51200,1.3410711288452148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,51200,0.5277768770853678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,16384,0.9819537268744575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,256,0.053992887338002525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,7168,0.136245330174764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,16384,0.02204799983236525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,64,0.013573333621025085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,16384,1.1615511576334636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,51200,0.11075644360648261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,32,0.012608889076444836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,65536,0.0765742195977105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,512,0.006115555349323485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,51200,0.03994489047262404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,2048,0.025647999511824712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,256,0.0035093331502543557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,3072,0.16021154986487493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,5120,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,2048,0.01867377758026123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,64,0.011967110964987012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,512,0.010337777435779572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,256,0.007344000041484833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,7168,0.023967110448413428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,7168,0.057408001687791615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,2560,0.00574311117331187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,16384,0.03179377648565505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,1536,0.18256799379984537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,1536,0.024921778175565932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,51200,0.18496444490220812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,3584,0.007265778051482306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,1024,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,16384,0.0614195532268948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,128,0.01200355589389801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,768,0.003801777958869934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,512,0.00444533344772127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,1024,0.009266667068004608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,3584,0.09513600005043878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,2560,0.05048355460166931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,10240,0.015505777464972602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,128,0.006786666810512543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,4096,0.006361777583758037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,4096,0.01163822246922387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,65536,0.18454221884409586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,3584,0.09394933117760552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,128,0.005892444401979446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,3584,0.2701280117034912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,16384,0.017645332548353408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,5120,0.012355555262830524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,51200,0.05312533179918925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,10240,0.0064693333374129395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,4096,0.00959644466638565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,16384,0.16657243834601507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,51200,0.11016711261537339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,2048,0.04274133178922865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,32,0.05606044663323296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,256,0.005038222091065513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,64,0.003104000041882197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,7168,0.015034667319721647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,3584,0.007878222399287755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,8192,0.029429333077536687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,3584,0.06001955270767212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,3584,0.0634746683968438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,128,0.019827556278970506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,1536,0.022037333912319605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,3584,0.0704631143146091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,10240,0.11760266621907552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,2048,0.007093333535724216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,512,0.0034595556143257353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,16384,0.10628088977601792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,51200,0.1797093285454644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,1024,0.008868444297048781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,2048,0.16617155075073242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,10240,0.016598222984208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,6144,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,10240,0.06232800086339315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,65536,0.5105306837293837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,51200,0.7726524670918783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,256,0.003830222205983268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,6144,0.010116444693671333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,1024,0.006335999816656113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,4096,0.010905777414639791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,8192,0.03823555509249369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,8192,0.02407288882467482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,128,0.004609777695602841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,5120,0.008772444393899705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,4096,0.01351288871632682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,2048,0.016524444023768108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,32,0.00518844442235099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,8192,0.057224002149369985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,10240,0.18817599614461264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,3072,0.012916444076432122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,10240,0.11578933397928874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,768,0.005691555639108022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,2560,0.023891554938422307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,65536,0.0691244469748603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,2560,0.008007999923494127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,64,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,6144,0.09498933288786147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,3072,0.08740088674757217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,1536,0.011873777541849347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,256,0.003742222156789568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,1536,0.017190222938855488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,3584,0.02689244515366024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,1024,0.03849422269397312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,10240,0.13791467083825007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,768,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,2048,0.5810640123155382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,1536,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,256,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,3584,1.088205337524414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,51200,0.05398844348059761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,65536,1.8235865698920355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,6144,0.015441776977645027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,8192,0.02204533252451155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,1024,0.0352497763103909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,12288,0.03457066747877333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,1536,0.11472888787587483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,5120,0.01389955480893453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,128,0.013251555462678274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,3584,0.021406221720907424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,5120,0.04095822241571214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,1024,0.004080888711743885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,128,0.015321777926550971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,32,0.0052177777720822235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,10240,0.15455378426445857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,5120,0.009454222189055549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,256,0.005669333454635408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,51200,0.4782240125868056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,10240,0.7137146525912814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,2048,0.008963555925422246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,6144,0.008207999997668797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,512,0.006033777776691649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,3584,0.021007110675175984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,5120,0.021612443857722815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,8192,0.008656000097592672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,512,0.007984889050324758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,5120,1.6543635262383356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,1024,0.4884595341152615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,128,0.013627555635240344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,256,0.005924444645643234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,7168,0.02029333346419864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,3584,0.9743742412990994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,2560,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,65536,0.442446231842041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,51200,3.430839114718967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,12288,0.46639733844333225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,5120,0.23738310072157118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,7168,0.011680888632933298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,2048,0.03457422057787577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,3072,0.007664889097213745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,2560,0.013078221844302284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,6144,0.021524444222450256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,10240,0.02313866714636485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,4096,0.07129066520267062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,512,0.0052986666560173035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,256,0.01402311192618476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,1536,0.0803297758102417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,16384,0.053744001521004565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,2048,0.013520888984203339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,8192,4.051152971055773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,3584,0.015998222761683993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,1536,0.012717333104875354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,768,0.016326222154829238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,51200,0.1198542250527276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,3584,0.00796977761718962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,10240,0.16689156161414254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,3584,0.008568888737095727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,2560,0.1524275541305542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,4096,0.022092445029152766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,2048,0.04792177677154541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,10240,0.0446320010556115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,2048,0.034865776697794594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,512,0.005468444277842839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,7168,1.8262799580891926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,1536,0.04571022258864509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,768,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,6144,0.19830932882097033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,2560,0.011348444554540845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,2560,0.015057777365048727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,10240,0.08539555470148723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,256,0.004873777843183941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,51200,6.837240854899089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,1536,0.02145066691769494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,5120,0.34396889474656844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,128,0.007162666983074612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,4096,0.5957075754801432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,1536,0.15301511022779676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,3072,0.0439662237962087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,65536,345.8948025173611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,65536,5.513715532090928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,7168,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,7168,0.1109386682510376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,51200,0.09317244423760308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,512,0.22914221551683214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,1024,0.016511110795868766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,1024,0.015676443775494892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,16384,0.1144791046778361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,8192,0.02346666653951009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,2048,0.01149511088927587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,2048,0.020854221449957952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,10240,0.15388888782925075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,2560,0.015480000111791821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,768,0.03695644603835212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,7168,1.7612471050686305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,4096,0.01129155523247189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,2048,0.009475555684831407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,32,0.014521777629852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,10240,0.0074622225430276655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,1024,0.032207111517588295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,3584,0.006701333241330252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,512,0.007603555917739868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,6144,0.01591822173860338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,5120,0.6444062127007378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,2048,0.02422222163942125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,51200,0.4959431224399143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,16384,0.0210533340771993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,768,0.008572444319725037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,7168,0.01960444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,1024,0.05085244443681505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,2048,0.01891377733813392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,4096,0.014682667122946845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,3072,0.00813422186507119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,10240,0.30377422438727486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,6144,0.011639110743999481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,51200,1.1863075892130535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,2048,0.02235377828280131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,65536,0.1125955581665039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,3584,0.023708444502618577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,768,0.005102222164471944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,128,0.013608000344700284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,256,0.00305244450767835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,768,0.01144444445768992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,1536,0.04775822162628174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,12288,0.015441776977645027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,1536,0.008801778157552084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,65536,0.13917511039310032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,4096,0.01976444489426083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,2048,0.014097778333557976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,3072,0.006353777729802662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,256,0.0046826667255825466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,51200,0.22543109787835014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,2048,0.007351111206743453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,2048,0.024118221468395654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,51200,0.16523644659254286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,51200,1.669991069369846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,2048,0.05785244703292847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,128,0.005709333138333426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,3072,0.03375466664632162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,256,0.004769777672158347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,3584,0.011591110792424945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,32,0.003206222214632564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,256,0.004398222184843487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,64,0.006504000061088138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,2560,0.9334071477254232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,1536,0.046816888782713145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,12288,0.04084622197681003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,32,0.01869155632125007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,6144,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,256,0.0030826667530669104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,65536,1.4457129372490778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,256,0.0035706667436493766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,6144,0.04714577727847629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,3584,0.0119991112086508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,512,0.037661333878835045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,5120,0.035876442988713585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,768,0.009015111459626092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,3072,0.010151999692122141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,65536,1.766154713100857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,128,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,6144,0.01221866657336553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,4096,0.028343998723559912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,10240,0.016559999850061204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,512,0.005561777700980504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,128,0.005063111169470681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,7168,0.030460443761613634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,65536,0.1496986680560642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,16384,0.09796000189251369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,128,0.0031235555393828284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,128,0.09111999803119236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,256,0.007859555383523306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,256,0.15328355630238852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,7168,0.1026115549935235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,3584,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,256,0.03645777702331543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,2048,0.9319093492296008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,12288,0.04047111007902358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,3584,0.01236533373594284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,2560,0.01953155630164676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,7168,0.04789333211051094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,512,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,4096,0.008157333566082848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,16384,0.3086835543314616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,51200,0.16577155060238308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,51200,0.21954845057593453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,16384,0.05318399932649401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,128,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,51200,0.09045955869886611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,2048,0.00905511114332411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,1024,0.006545777950021956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,1536,0.035529779063330755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,3072,0.01090755561987559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,64,0.0027599999060233435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,128,0.004039111236731212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,128,0.0031848889258172777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,1024,0.0148062225845125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,6144,0.1341537766986423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,3072,0.18378844526078966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,6144,0.03724088933732774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,64,0.002672000063790215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,3584,0.05614666806326973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,51200,0.4559111065334744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,8192,0.47136534584893125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,7168,0.47549067603217227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,1536,0.10911911063724095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,2048,0.06101510922114054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,3584,0.3208844396803114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,768,0.011379555695586734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,1536,0.009026666482289633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,8192,0.022745778163274128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,16384,0.22381067276000977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,1024,0.00870577742656072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,1024,0.005707555760939916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,4096,0.0396106673611535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,3584,0.009033777647548253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,32,0.003040888864133093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,2560,0.5521173477172852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,65536,0.12806577152676052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,2048,0.011287111375066968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,8192,0.02332088847955068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,2560,0.008240000241332585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,6144,0.08944177627563477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,1536,0.023779556155204773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,512,0.006540444576077991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,6144,0.035938666926489934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,6144,0.005935110979610019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,5120,0.1677537759145101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,512,0.00424799985355801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,12288,0.10158133506774902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,512,0.2582204341888428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,12288,0.03631733192337884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,8192,0.007429333196745978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,12288,0.6672177844577365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,256,0.009487999810112847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,2560,0.007110222346252865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,5120,0.011936888926559024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,512,0.0359226663907369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,1536,0.02085777786042955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,5120,0.030001779397328694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,7168,1.1059012942843967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,64,0.005787555542257097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,65536,91.41626654730902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,3072,0.018613333503405254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,512,0.01642666591538323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,1536,0.00961866643693712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,1024,0.038359112209743924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,65536,0.09153866767883301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,16384,0.17547289530436197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,2048,0.03894666830698649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,4096,0.015435554915004306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,16384,0.0083137775460879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,4096,0.030858665704727173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,2048,0.012549333274364471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,65536,0.07447999715805054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,64,0.0025920000755124623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,128,0.003220444545149803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,768,0.009930666950013902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,768,0.005164444446563721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,768,0.008658666577604082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,7168,0.07574311229917738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,768,0.01404800017674764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,8192,0.01442577772670322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,768,0.012221333053376941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,16384,0.037541333172056414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,32,0.005175999883148405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,2560,0.9440577824910482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,16384,0.17749867174360487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,10240,0.018831999765502084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,51200,4.2878846062554254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,64,0.01070577816830741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,64,0.004107555581463708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,2048,0.016540444559521146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,128,0.0032471112079090546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,512,0.013363555901580386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,16384,0.5607982211642796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,4096,0.0069439998931354946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,3584,0.024643555283546448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,1536,0.029704888661702473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,5120,0.021054221524132624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,10240,0.4769146707322862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,16384,0.08972355392244126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,65536,0.08079289065466987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,256,0.004126222183307012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,2560,0.24473865826924643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,10240,0.03258400162061056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,10240,0.01423288881778717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,128,0.012841777669058906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,1536,0.006622222148709827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,4096,0.0105253333846728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,256,0.0046364445653226645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,3072,0.04513600137498644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,65536,0.12483821974860297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,5120,0.011423110961914062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,256,0.003132444495956103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,4096,0.01629244453377194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,4096,0.016190222567982145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,768,0.04180533356136746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,4096,0.01201777739657296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,512,0.0041164445380369825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,51200,7.1754650539822045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,32,0.0038133333954546186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,64,0.0075937774446275495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,128,0.010272888673676385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,7168,0.04843466811709934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,10240,1.1092684004041884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,4096,0.011209777659840055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,1536,0.09481333361731635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,3584,0.010509333676762052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,5120,0.03812177644835578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,1024,0.00544799988468488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,128,0.008643555144468943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,5120,0.008360889222886827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,1024,0.007167110840479533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,3072,2.072164535522461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,768,0.021820444199774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,128,0.04548266530036926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,3072,0.045532445112864174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,2048,0.04091733362939622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,1024,0.08623733123143514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,1536,0.01201244443655014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,8192,0.013159111142158508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,7168,0.09017778105205959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,16384,0.10332710875405206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,12288,0.02796622117360433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,512,0.004949333353175057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,3072,0.07129866547054715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,7168,0.8461084365844727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,256,0.005106666849719154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,64,0.006531555619504716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,16384,0.04916533496644762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,16384,0.5374551349216038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,65536,2.481449762980143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,7168,0.04541244440608554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,10240,0.017754667335086398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,3584,0.06381689177619086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,1536,0.0419680012596978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,4096,0.03165955675972833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,51200,0.16273778014712864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,65536,0.19973511166042754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,1024,0.018238221605618794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,128,0.009660444325870937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,65536,2.417427486843533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,2048,0.014310222533014087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,128,0.005072888814740711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,128,0.03200622068511115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,51200,0.062272886435190834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,768,0.022602667411168415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,10240,0.7879075474209256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,256,0.01497066683239407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,65536,0.46332444085015195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,2048,0.0069644442862934535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,5120,0.026960889498392742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,3072,0.008641777767075432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,64,0.009008888569143083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,256,0.015406222807036506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,32,0.0032008888406885993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,256,0.004613333278232151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,1024,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,64,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,16384,0.05911111169391208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,768,0.005199999858935674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,3072,0.007064888874689738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,2048,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,12288,0.022031999296612207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,6144,0.06696977880265978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,3072,0.27964621120029026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,8192,0.011759999725553723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,16384,0.019278221660190158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,7168,0.08166488673951891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,2048,0.0088755554623074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,7168,0.02896089024013943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,8192,0.038021332687801786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,51200,1.3355422549777562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,4096,0.3389093346065945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,65536,0.022143110632896423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,8192,3.3497831556532116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,7168,6.798026614718967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,4096,0.9334328969319662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,32,0.0032275555034478507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,10240,0.07847555478413899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,4096,0.05111289024353027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,5120,0.10581333107418484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,2048,0.03249866763750712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,8192,7.096015082465278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,12288,0.3722240130106608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,2560,0.12734133667416042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,16384,0.49741776784261066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,51200,0.17380089230007598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,128,0.009513777991135916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,51200,1.9178791046142578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,6144,0.07959466510348849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,10240,4.572572496202257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,5120,0.010445333189434476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,1024,0.01148799972401725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,128,0.003949333396222857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,4096,1.4962729348076715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,2048,0.010481778118345471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,768,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,3072,0.009693333672152625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,4096,0.014010666145218743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,8192,0.05297777718967862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,2560,0.0417902237839169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,768,0.0037084445357322693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,65536,0.13696711593204075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,1536,0.006520888871616787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,12288,0.5426755481296116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,7168,0.038246221012539335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,512,0.08461688624487983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,512,0.1539751158820258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,6144,0.017089777522616915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,1024,0.015146666102939181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,7168,2.4542577531602645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,7168,0.021632888250880774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,51200,0.06643466817008124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,7168,0.4484657711452908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,3584,0.08056710826026069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,256,0.0036311112344264984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,512,0.00906577789121204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,1536,0.011741333537631564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,3072,0.006762666834725275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,12288,0.15871200296613905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,51200,0.12176888518863255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,8192,0.0820373362965054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,12288,0.08732799688975017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,256,0.0063635553750726916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,7168,0.013574221895800697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,3584,0.009215110705958473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,1536,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,51200,0.14376800590091282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,2560,0.039176887936062284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,51200,0.05380622214741177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,3584,0.0184888889392217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,7168,0.020791111720932853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,4096,2.445869233873155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,65536,0.31025777922736275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,7168,0.04087555408477783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,256,0.00794933322403166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,51200,0.09450311130947536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,65536,1.8371910519070094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,768,0.023364444573720295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,1024,0.07492088609271579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,2048,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,512,0.003921777837806278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,51200,0.061927113268110484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,16384,0.2859128846062554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,512,0.006072000082996156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,256,0.003132444495956103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,10240,0.20194044378068712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,4096,0.11312977472941081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,4096,0.010234666367371878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,1024,0.005657777604129579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,1024,0.004109333372778363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,51200,0.16473600599500868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,3072,0.10221511125564575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,6144,0.006344888773229387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,1536,0.11948088804880778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,16384,4.507503085666232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,8192,0.01891911029815674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,5120,0.013178666432698568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,16384,3.034237331814236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,5120,0.20041067070431182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,512,0.006599999964237213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,1536,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,128,0.0036319999231232535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,12288,0.28387376997205943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,64,0.007189333438873291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,128,0.010072888599501716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,128,0.0029306666304667792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,3584,0.0424542228380839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,3072,0.00923466682434082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,512,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,128,0.007807999849319458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,16384,0.027426666683620874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,51200,4.0121714274088545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,128,0.04323022233115303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,2048,0.06857333580652873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,3584,0.06255733304553561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,3072,0.008423111504978603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,10240,0.057421333260006376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,10240,0.039699554443359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,8192,0.014460444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,51200,0.12109955151875813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,768,0.0058293334311909145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,512,0.02110222147570716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,1536,0.012890666723251343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,10240,0.039537777503331505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,3072,0.08181777927610609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,12288,0.019677332705921598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,16384,0.06499466631147596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,2048,0.015862221519152325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,32,0.009696889254781935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,3584,0.008829333715968663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,64,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,3584,0.01737511157989502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,128,0.026940445105234783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,7168,1.9679679870605469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,5120,0.04749333196216159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,512,0.004639111045334074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,1024,0.045567999283472695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,16384,0.025435555312368605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,1536,0.0069848886794514125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,6144,0.019739554988013375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,32,0.0033288889875014624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,10240,0.37955644395616317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,2048,0.054570668273501925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,2560,0.009533333281675974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,8192,0.029924445682101782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,1024,0.016910221841600206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,4096,0.02549244463443756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,3584,0.06830933358934191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,10240,0.08677955468495686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,7168,0.3263431125217014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,4096,0.006162666612201267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,10240,0.19576799869537354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,3072,0.006636444479227066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,65536,0.67922666337755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,1536,0.09412888685862224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,32,0.01387733303838306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,768,0.04489244355095757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,768,0.0537431107627021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,128,0.011776888536082374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,10240,0.11060444513956706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,64,0.002712888850106133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,12288,0.026870222555266485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,6144,0.095432890786065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,768,0.003702222059170405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,1024,0.008547555241319869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,12288,0.02123733361562093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,1536,0.010045333041085137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,12288,0.13781066735585532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,51200,0.601861317952474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,128,0.003591999825504091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,128,0.0029120000286234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,6144,0.013295110729005603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,3072,0.007313778003056844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,7168,0.17653332816229927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,3072,0.008720888859695857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,512,0.0035502223504914176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,16384,0.020977778567208186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,4096,0.04291911257637871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,256,0.005487999982304043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,16384,0.5245555771721734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,12288,0.05614933040406969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,10240,0.13369866212209067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,12288,4.486402723524305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,5120,0.09555200073454116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,16384,0.026512889398468867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,65536,0.17795466052161324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,6144,0.046591109699673124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,2560,0.014215999179416232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,32,0.0033128888656695685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,7168,0.09470933013492161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,64,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,51200,0.09003733264075385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,5120,0.09374844365649754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,1536,0.015449777245521545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,32,0.0036684444381131064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,8192,0.013929777675204806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,8192,0.02215733296341366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,7168,0.09648711151546902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,3072,0.015455111861228943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,16384,12.695694817437065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,512,0.005579555614127054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,128,0.004387555436955558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,6144,0.017808889349301655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,2560,0.011324444578753578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,1536,0.007855999800893996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,1024,0.017835555805100333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,512,0.003903111235962974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,256,0.003154666680428717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,1536,0.038591110044055514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,4096,0.008824889030721452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,256,0.04263910982343885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,4096,0.018570666511853535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,65536,0.2665857738918728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,2560,0.00775822252035141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,512,0.005312888986534542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,5120,0.09082488881217109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,1536,0.01089511149459415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,1536,0.08366311258739895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,51200,1.6148018307156031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,128,0.00813244448767768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,51200,1.3350000381469727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,16384,0.05390933487150404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,8192,0.01794933279355367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,128,0.004468444320890638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,6144,0.029891557163662378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,4096,0.028723554478751287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,2048,0.01276622215906779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,12288,0.006050666587220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,3584,0.007984889050324758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,2560,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,12288,0.08984266387091742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,512,0.04202666547563341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,10240,0.4330337842305501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,1536,0.018418666389253404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,10240,0.02277244461907281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,8192,0.1808204385969374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,64,0.005507555686765247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,16384,0.34270577960544163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,64,0.004129777765936322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,12288,0.05332622263166639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,7168,0.030590222941504583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,1024,0.05579555696911282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,2048,0.0070017774899800616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,128,0.007154666715198093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,7168,2.085850609673394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,3584,0.03385244475470649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,3584,0.028220444917678833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,10240,9.280931260850695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,5120,0.016700444949997794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,16384,0.040961778826183744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,2560,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,2560,0.00609422226746877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,3072,0.007698666718271043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,2048,0.0173662222094006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,4096,0.06703822480307685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,1536,0.00700711127784517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,256,0.007829333345095316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,7168,0.012893333203262754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,65536,0.07258933120303683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,65536,2.1983636220296225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,512,0.005694222119119432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,65536,0.8047093285454644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,128,0.00981866651111179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,512,0.009604444106419882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,1024,0.013535111314720578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,51200,0.049829334020614624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,3584,0.1341866652170817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,6144,0.017328000730938382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,256,0.02715822226471371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,256,0.0029475556479560006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,128,0.011936888926559024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,16384,0.0677973363134596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,1024,0.00610133343272739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,32,0.008074666890833113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,4096,0.01973155637582143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,4096,0.2274951140085856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,2048,0.05353600117895338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,51200,2.562893337673611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,1024,0.02777155571513706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,512,0.010933332973056369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,6144,0.06462488571802776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,51200,0.24826576974656847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,128,0.0027422221998373666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,5120,0.013719999955760108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,1536,0.010092444717884064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,65536,1.2116115358140733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,5120,0.01867377758026123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,16384,0.06945688856972589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,3584,0.14050843980577257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,10240,0.03757866554790073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,1024,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,8192,0.26403821839226616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,8192,0.02889333168665568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,64,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,7168,0.01958400011062622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,256,0.009552888572216034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,10240,0.36124711566501194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,4096,0.7746417787339953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,3072,0.008887999587588841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,512,0.004506666627195147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,7168,0.01797155539194743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,128,0.015936000479592215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,8192,0.18789154953426787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,10240,0.013448889056841532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,768,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,5120,0.2843155595991346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,65536,0.21069867081112334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,128,0.014045332868893942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,64,0.006107555495368109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,512,0.00463377767139011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,512,0.023530667026837666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,768,0.009189333352777693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,3584,0.11513777573903401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,1536,0.027552889453040227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,8192,0.4427608913845486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,3584,0.010671111444632212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,8192,0.6859075758192273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,32,0.003124444435040156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,5120,0.01314399970902337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,256,0.005688000056478713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,64,0.00638133328821924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,3584,0.027293332748942908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,8192,0.026913776993751526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,768,0.1032231118943956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,3072,0.11438488960266113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,5120,0.011662222445011139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,1024,0.016359999775886536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,768,0.004421333471934001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,3584,0.08561066786448161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,16384,0.05528266562355889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,65536,1.9981822967529297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,8192,0.015461333923869662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,2560,0.010790222220950656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,4096,0.017765333255132038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,16384,0.2153306669659085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,8192,0.0092622223827574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,7168,0.026408000124825373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,51200,39.28643459743923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,8192,0.020029332902696397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,1536,0.0092613332801395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,7168,0.008918222453859117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,51200,0.21004888746473524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,8192,0.8055235544840494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,7168,0.34348177909851074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,2048,0.014866666661368476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,10240,0.04658311274316576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,16384,0.1089751124382019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,12288,0.18826756212446424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,16384,0.20045777161916098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,64,0.19669066535101998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,5120,0.020495999190542433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,3584,0.09174488650427924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,64,0.009327111144860586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,65536,0.13694222768147787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,3584,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,32,0.021293333835071985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,1024,0.48760800891452366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,16384,1.117221302456326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,65536,0.06689155764049955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,8192,0.04303110970391167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,7168,0.022944889134830896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,1024,0.03461066550678677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,8192,0.2834248807695177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,64,0.01646133263905843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,12288,0.07814221911960177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,10240,0.09793600108888413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,768,0.0354639987150828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,3072,0.014726221561431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,3072,0.07721955246395536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,65536,0.07725510994593303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,7168,0.015056000815497505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,2048,0.06103822257783678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,10240,0.07957600222693549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,1024,0.006810666786299811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,6144,0.18237333827548555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,5120,1.264269298977322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,3584,0.06565688716040717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,6144,6.169832017686631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,3584,0.33842934502495664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,12288,0.024040000306235418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,2048,0.008811555802822113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,10240,0.4907289081149631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,4096,0.3049235608842638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,32,0.02625777820746104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,7168,0.04723911152945625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,32,0.007479111353556315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,1536,0.0313662224345737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,2560,0.13410133785671657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,256,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,16384,0.13329154915279812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,4096,0.07698311408360799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,1536,0.0119982221060329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,6144,1.2389386494954426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,3584,0.035325331820382014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,1024,0.07485244671503703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,1536,0.005502222312821283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,1536,0.013626666532622443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,32,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,64,0.008616888688670265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,512,0.005600888695981767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,768,0.010281778044170803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,8192,0.01569600072171953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,128,0.0033982222278912864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,64,0.06941866874694824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,2560,0.435315555996365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,256,0.007638221813572778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,256,0.00442755553457472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,128,0.006266666783226862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,2560,0.30607554647657603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,6144,0.21127733919355604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,16384,2.282525380452474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,12288,0.1262622276941935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,51200,0.15402844217088488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,64,0.007786666353543599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,12288,0.021194666624069214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,1024,0.025249777568711176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,4096,0.02232355541653103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,5120,0.018062222335073683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,512,0.04723377691374885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,3072,0.01092088884777493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,1024,0.015105777316623263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,12288,0.019850666324297588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,65536,0.07027200195524427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,10240,1.3299858305189345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,16384,0.26291023360358345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,64,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,128,0.004795555439260271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,512,0.029952890343136255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,2048,0.021464889248212177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,2048,0.0377786656220754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,3072,0.08462133010228474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,2048,0.14487111568450928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,2560,0.005972444597217772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,256,0.007311111523045435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,256,0.007802666889296637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,512,0.3215013345082601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,4096,0.013614222407341003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,256,0.052525334888034396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,512,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,768,0.08519466718037923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,2560,0.26120267974005806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,12288,0.05485777722464668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,512,0.01293777757220798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,512,0.006024888820118374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,512,0.009335111412737105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,64,0.002512000087234709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,3584,0.03237688872549269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,8192,0.05434222353829277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,12288,0.03674933314323425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,256,0.005693333430422678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,12288,0.09579110807842678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,5120,0.013162666724787818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,512,0.005720888988839255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,8192,0.02176622218555874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,1536,0.061140444543626576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,65536,0.14955911371443006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,10240,0.13695555263095432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,2560,0.01901333365175459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,128,0.06555200285381742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,512,0.015240889456537036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,51200,0.9639608595106336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,65536,0.11132621765136719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,4096,0.009104000197516548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,512,0.006638222270541721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,7168,0.037676443656285606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,512,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,256,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,1024,0.003964444415436851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,10240,0.02239022155602773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,1536,0.04154310954941644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,256,0.031046221653620403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,65536,0.4375217755635579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,64,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,1024,0.004805333498451444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,1024,0.011396444506115384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,1536,0.06775110960006714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,256,0.03766311208407084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,16384,0.036814222733179726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,3584,0.017383111847771537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,65536,3.3748346964518228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,2048,0.11208000448015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,10240,2.2658053504096136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,6144,0.2418417665693495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,2560,0.00999466660949919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,5120,0.014789332946141561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,3584,0.012479999827014076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,4096,0.2346737649705675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,4096,0.008408889174461365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,12288,0.19140444861518013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,128,0.014517333772447375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,2560,0.024665777881940205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,2048,0.04028622309366862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,1024,0.007925333248244392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,64,0.0032337777730491427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,512,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,5120,0.15483466784159342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,51200,0.012489777472284106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,51200,0.3574364450242784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,2048,0.11870488855573867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,8192,0.024541333317756653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,32,0.0027537778433826235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,256,0.014107555150985718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,32,0.00777511133088006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,65536,0.424998230404324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,1024,0.005752000129885144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,1536,0.006435555716355641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,5120,0.027544000082545813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,1536,0.053986665275361806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,3584,0.0904373327891032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,7168,0.03840622305870056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,2048,0.27306310335795086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,2560,0.011962666279739805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,1024,0.009533333281675974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,3072,0.010758221977286868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,4096,0.056571556462181934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,2048,0.01202400028705597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,32,0.16896977689531115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,2560,0.007442666424645319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,5120,0.008952889177534316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,256,0.011982222398122152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,51200,0.06548888815773858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,65536,1.6131750742594402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,51200,15.861609564887154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,12288,4.653824700249566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,4096,0.05007911059591505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,51200,4.019993676079644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,7168,0.03323910964859857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,16384,0.06283999813927545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,2048,0.7022844420539008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,6144,0.024668445189793903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,3072,0.010071111222108206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,6144,0.022220444348123338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,1536,0.006600889066855113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,768,0.006574222197135289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,4096,0.08716622326109146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,6144,0.011401777466138205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,768,0.011559999651379056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,4096,0.012660444610648684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,5120,0.09964178005854289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,32,0.00498577786816491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,5120,0.01776622235774994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,3584,0.018562666243977018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,768,0.02907199992073907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,16384,0.40729689598083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,1024,0.9648222393459744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,65536,0.0795653329955207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,65536,0.5852648946974013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,10240,0.272145774629381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,10240,0.014721777704026965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,128,0.0029502221279674103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,16384,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,65536,0.30711910459730357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,3584,0.018567999203999836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,2560,0.01936177743805779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,128,0.0041831110914548235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,3072,0.09331111113230388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,8192,1.1830515331692166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,4096,0.10773689217037624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,2560,0.05986933575736152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,2560,0.05372977919048733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,10240,0.3613573445214166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,16384,0.056559999783833824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,10240,0.015454222758611044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,3072,0.08740533060497707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,51200,0.08248088757197063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,3584,0.023103111320071753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,4096,1.8652328915066188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,3584,2.622901280721029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,128,0.030689779255125258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,256,0.004127999974621667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,2560,0.07556266917122735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,32,0.0031991110493739447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,5120,2.260801739162869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,3584,0.22061688370174834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,6144,0.020384889509942796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,8192,0.025143999192449782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,256,0.0038399998512532976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,2048,0.1293520000245836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,256,0.005067555440796746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,2048,0.042118221521377563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,6144,0.021149333980348375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,12288,0.813044442070855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,4096,0.015736889508035447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,12288,0.06298844681845771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,8192,0.007753777835104201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,6144,0.03324177861213684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,256,0.004383999854326248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,10240,0.3739679919348822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,3072,0.09901955392625596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,4096,1.2948764165242512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,10240,0.02604088849491543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,7168,1.577736006842719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,1024,0.0075911109646161394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,1024,0.05050666795836555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,8192,2.9354915618896484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,1536,0.02544622288809882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,65536,0.4020435545179579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,7168,0.013200889031092325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,3584,0.12907200389438206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,1024,0.01149866647190518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,51200,0.11696088314056396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,5120,0.02203911046187083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,3584,0.008051555189821456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,7168,0.07644355297088623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,256,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,51200,0.11714222696092393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,10240,0.051764445172415845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,256,0.036166220903396606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,2560,0.008410666551854875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,10240,0.0398613346947564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,10240,0.01368799971209632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,51200,2.084399117363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,2560,0.010888000329335531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,6144,0.018277333842383493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,128,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,3072,0.011567110816637674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,2048,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,4096,0.026351110802756414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,1536,0.02880711025661892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,128,0.18178044425116646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,128,0.008115555677149031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,3072,0.014327999618318347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,1536,0.035367111365000405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,768,0.008841777841250101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,256,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,12288,5.767689598931207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,7168,0.03827733463711209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,1024,0.19804978370666504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,256,0.00700800038046307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,128,0.006527110934257507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,16384,0.05656444364123874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,768,0.04576711191071404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,1024,0.006879999819729064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,128,0.008081778056091731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,256,0.005683555371231503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,768,0.019638222124841478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,12288,0.0959671139717102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,256,0.14386933379703098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,65536,0.3879084322187636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,256,0.013814222481515674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,1024,0.025535111625989277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,512,0.004913777940803104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,64,0.006993778049945831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,51200,2.3401972452799478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,12288,0.09040000041325887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,512,0.008431110945012834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,3584,0.007136888802051544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,51200,0.090865777598487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,12288,0.04308888978428311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,128,0.03465866711404588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,1024,0.007426666716734569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,2048,0.011191111471917896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,256,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,3584,0.21926400396558973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,768,0.023199111223220825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,12288,0.0403386652469635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,2048,0.10157777865727742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,16384,0.36108356051974827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,65536,0.07501333289676242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,1536,0.067266669538286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,65536,0.015111999379263984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,12288,0.03231733375125461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,6144,0.025195555554495916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,7168,2.14628537495931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,1024,0.04577066832118564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,10240,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,12288,0.04777333471510145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,5120,0.03698044353061252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,4096,0.11601689126756455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,32,0.003039999968475766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,1024,0.006102222121424145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,1024,0.029049777322345312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,2048,0.006759111252095964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,7168,0.04766133427619934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,1024,0.012496888637542725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,12288,0.21135377883911133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,8192,0.17452355225880942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,256,0.005149333427349727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,4096,0.0083155557513237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,3584,0.008946666287051307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,7168,0.027277333868874445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,128,0.009413333402739631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,4096,0.03538399934768677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,1536,0.01330044451687071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,10240,0.013602666556835175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,2048,0.013932444983058505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,1536,0.004707555390066571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,1024,0.09968888759613037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,64,0.003101333354910215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,2560,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,1024,0.004884444591071871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,1536,0.01904444396495819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,51200,0.5205698013305664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,1024,0.050752888123194374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,2048,0.06973510980606079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,51200,0.882023122575548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,10240,0.06528977553049724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,7168,0.0138586668504609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,1024,0.028869334194395278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,2560,0.0210524449745814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,3072,0.06225155459509956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,16384,3.646386464436849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,16384,0.31828000810411244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,1536,0.05164177881346809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,10240,0.2344435585869683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,4096,0.140864888827006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,4096,0.013983110586802164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,768,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,16384,0.12195200390285915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,128,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,128,0.003112000102798144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,64,0.002663111107216941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,5120,0.011945777469211154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,768,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,768,0.010252444280518426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,32,0.006048888795905643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,2048,0.012088889049159156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,1024,0.021272889441914026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,5120,0.017843556072976854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,1024,0.04008355405595567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,4096,0.1352355612648858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,768,0.00646666685740153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,256,0.005649777750174205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,2560,0.009082666701740688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,3072,0.4047039879692926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,768,0.0650435553656684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,12288,0.04667822188801236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,768,0.006970666348934174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,4096,0.2590888871086968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,65536,0.18948977523379854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,65536,0.25934134589301217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,12288,0.06676355335447523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,1536,0.010536889235178629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,1536,0.008428444465001425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,4096,0.03880000114440918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,512,0.005582222094138463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,2560,0.018910222583346896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,5120,0.008802666432327693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,768,0.009656888743241629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,7168,0.05836088789833916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,128,0.007141333487298753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,2048,0.00590133335855272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,5120,0.20871110757191977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,51200,0.7385039859347873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,4096,0.06936799817615084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,16384,6.202745649549697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,4096,0.03718577822049459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,51200,0.12317332956526016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,7168,0.06714755296707153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,8192,0.4083555539449056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,128,0.012083555261294046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,2560,0.00638755535085996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,16384,2.406735102335612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,3072,0.01333866682317522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,32,0.0064497776329517365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,768,0.021202666891945735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,65536,1.9903182983398438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,2048,0.09729155566957261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,51200,0.09107110897699992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,5120,0.14184533225165472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,512,0.009864889085292816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,65536,0.2165520058737861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,3584,0.14391288492414686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,1536,0.07181155681610107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,3072,0.24202489852905273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,3584,0.023001778456899855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,1536,0.009031111167536842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,12288,0.26655644840664333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,64,0.0052435555391841466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,64,0.015388444066047668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,3072,0.016122667325867545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,65536,22.558485243055557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,5120,0.05662755833731758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,512,0.006169777777459886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,1536,0.00851288851764467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,5120,1.4419671164618597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,2048,0.018596443865034316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,32,0.003875555677546395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,1536,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,512,0.004972444640265571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,10240,0.017151110702090792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,1536,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,8192,0.01975200076897939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,10240,0.7928373548719617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,1024,0.01086400035354826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,512,0.04152977797720168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,1024,0.011311111350854238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,6144,0.028223110569847956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,2048,0.01923466722170512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,65536,0.12845333417256674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,65536,0.12354755401611328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,4096,0.23961512247721353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,2048,0.04690577917628818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,128,0.005132444616821078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,768,0.02905422117975023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,10240,0.521691534254286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,7168,0.5689457787407769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,65536,2.276086171468099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,51200,0.188119994269477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,65536,0.27569156222873265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,1536,0.04933511217435201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,1024,0.055895997418297656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,4096,0.18801955382029215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,10240,0.05884889099333021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,12288,0.016864889197879367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,16384,0.21607466538747153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,7168,0.03988622294531928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,4096,0.022783999641736347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,6144,0.0073520003093613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,512,0.021311110920376245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,32,0.02996088729964362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,64,0.009831111464235518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,1536,0.029501332177056208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,256,0.02386755579047733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,3584,0.010091555615266165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,3072,0.048466665877236255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,5120,0.03373688790533278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,12288,0.1472568909327189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,10240,0.015689777003394235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,32,0.021721776988771226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,12288,0.05744444661670261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,1024,0.005958222266700532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,768,0.016931555337376065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,4096,3.8880958557128906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,1536,0.04657422171698677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,51200,0.360302209854126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,768,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,768,0.010068444742096795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,1536,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,7168,0.04323200053638882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,4096,0.04312444395489163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,256,0.004381333374314838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,3584,0.008529778156015608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,128,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,65536,0.0400257772869534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,1024,0.022112889422310725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,3072,0.12870577971140543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,5120,0.014949333336618213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,512,0.06253511375851102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,128,0.006088888893524806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,768,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,5120,0.0177848885456721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,51200,0.008389333056079017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,512,0.005473777651786804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,32,0.0028231110837724474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,12288,0.03708088729116652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,65536,0.36096623208787704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,51200,0.06356800264782377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,7168,0.0214346663819419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,3072,0.05400088760587904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,65536,2.053568098280165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,5120,0.2314159870147705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,5120,0.05554044577810499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,1536,0.011445333560307821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,32,0.009184000392754873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,51200,0.05287733342912462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,2048,0.01862844493654039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,2560,0.30809688568115234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,16384,0.039065778255462646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,51200,22.7553473578559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,6144,0.24421244197421602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,768,0.13123822212219238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,2048,0.007532444265153673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,51200,0.03566666775279575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,256,0.011957333319716983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,128,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,1024,0.004049777984619141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,6144,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,32,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,12288,0.016414221790101793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,3584,0.5235413445366753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,768,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,1024,0.004009777886999978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,16384,0.04340799980693393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,3584,0.08227733108732435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,1536,0.009583111438486312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,64,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,65536,13.359019809299044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,65536,1.5392764409383137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,51200,0.06888533300823636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,3584,0.17260089185502794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,65536,3.0853173997667103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,1536,0.00868888861603207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,32,0.007855110698276097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,7168,0.19852977328830293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,128,0.046238223711649575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,1024,0.011396444506115384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,65536,0.7005075348748101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,12288,3.358794742160373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,256,0.004023111114899318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,128,0.003571555432346132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,64,0.006285333385070165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,16384,0.061463110976748996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,12288,0.05338489015897115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,10240,0.08008088668187459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,2560,0.01628444426589542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,7168,0.04780177937613594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,5120,0.2346728907691108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,1536,0.015768888923856948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,12288,0.019609777463806998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,1536,0.01804266704453362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,12288,0.03182222114668952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,1024,0.10636177990171646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,12288,1.6486843956841364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,5120,0.04158844550450643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,2048,0.03460888730155097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,1024,0.02481511069668664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,256,0.003056888985964987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,1024,0.010422222316265106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,2048,0.007349333001507654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,5120,0.016168889072206285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,65536,0.09178666936026679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,6144,0.015861334072219003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,65536,0.3860390981038411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,10240,0.13595822122361925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,16384,0.02456355591615041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,2560,0.023044443792766992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,3072,0.7188871171739367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,16384,1.8151599036322699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,2560,0.009203555683294931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,51200,0.07795466979344685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,5120,0.008555555509196388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,16384,0.025303999582926433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,2560,0.013537777794731988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,7168,2.5495368109809027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,256,0.004856889032655292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,3584,0.09954310788048638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,2560,0.04976266622543335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,2560,0.024879111184014216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,16384,0.027081777652104694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,65536,0.028767999675538804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,2048,0.01701955497264862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,1536,0.0074879998962084455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,3072,0.10075555245081584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,64,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,1024,0.01088711122671763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,12288,0.6543315781487359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,64,0.008554666406578487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,128,0.003699555579158995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,512,0.06658399767345853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,5120,0.012107555237081317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,128,0.011843555503421359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,12288,0.021789333886570398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,32,0.002625777696569761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,51200,0.7847866482204862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,3584,0.010410666465759277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,32,0.002830222249031067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,2048,0.006566222343179915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,2560,0.02426044477356805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,32,0.0030906666070222855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,8192,0.034690668185551964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,2048,0.01399466726515028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,12288,0.024725332856178284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,3072,0.48217423756917316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,256,0.005330666485759947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,6144,0.010950222611427307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,12288,0.029450668228997126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,7168,0.25223912133110893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,2560,0.14947911103566489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,2560,1.0525858137342665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,51200,0.07394755548901029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,10240,0.6456693543328179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,1024,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,64,0.0031831111345026228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,256,0.007800888684060838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,128,0.009776888622177972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,2048,0.03231466809908549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,6144,0.27053867446051705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,128,0.006487111250559489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,128,0.0033324443631701996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,2560,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,10240,0.02325155503220028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,64,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,128,0.014711999230914645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,512,0.05087022317780388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,6144,0.026879111925760906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,3584,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,51200,0.05883200301064385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,1536,0.004780444420046276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,768,0.006055111272467508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,8192,0.5798684226142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,5120,0.024554666545655992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,2048,0.0155404441886478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,1024,0.006822222222884496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,3584,0.019471999671724107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,12288,0.07904889186223348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,512,0.038627554972966514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,10240,0.08960977527830337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,3584,0.008586666650242275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,3584,0.027062222361564636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,8192,0.011668444507651858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,32,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,256,0.0029333333174387612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,3072,0.24907377031114367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,3584,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,256,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,1024,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,5120,0.01940088967482249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,10240,0.038069334295060896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,256,0.004382222063011593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,51200,0.24393778377109102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,2048,0.16534754965040419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,12288,0.1786968840493096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,8192,0.07296266820695665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,7168,0.18503110938602021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,32,0.09713244438171387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,2048,0.007326222128338284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,7168,0.011295999917719098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,2560,0.006092444476154115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,3072,0.01954844428433312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,6144,0.23122665617201063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,12288,0.14948799875047472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,1536,0.0269777774810791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,768,0.0387386679649353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,65536,2.6162747277153864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,128,0.004426666845877965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,65536,0.10301244258880615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,2560,0.20615911483764648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,128,0.01127377814716763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,51200,0.9807066387600369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,5120,0.1394595570034451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,12288,0.04879022306866116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,12288,0.040559109714296125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,32,0.017095999585257638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,6144,0.010078222387366826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,1536,0.00562755556570159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,7168,0.01240355521440506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,16384,0.562064011891683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,6144,0.01680266691578759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,51200,0.054238223367267184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,12288,0.022836445106400385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,768,0.051776889297697276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,768,0.008549333446555668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,64,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,2560,0.008433777425024245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,256,0.008307555483447181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,1024,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,2048,0.010760889285140567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,12288,0.025927111506462097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,768,0.004085333396991094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,16384,0.03183199961980184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,65536,1.7478026284111872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,256,0.0029208889851967492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,4096,0.4661102294921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,51200,0.33233245213826496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,2048,0.0247653325398763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,4096,0.12848711013793945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,7168,0.017584888471497428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,128,0.00434933313065105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,16384,0.023471999499532912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,256,0.04532266656557719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,1024,0.015791111522250705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,3072,0.012614222036467658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,128,0.04746133420202467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,2048,0.01348000019788742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,5120,0.052707556221220225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,6144,0.019336000084877014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,10240,0.19719110594855416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,768,0.0064284445510970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,4096,0.0525182220670912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,4096,0.04808533191680908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,1024,0.0057813334796163775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,12288,0.017507554756270517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,256,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,6144,0.07762489053938124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,4096,0.008430221842394935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,1024,0.0498542222711775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,256,0.016948444975747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,10240,0.18895288308461508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,1024,0.025819554924964905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,5120,0.060666667090521924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,7168,0.13837956057654485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,3584,0.011402666568756104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,2560,0.009552888572216034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,12288,0.0211191111140781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,1536,0.04756177796257866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,10240,0.03575377662976583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,12288,0.229327122370402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,128,0.024304888314670984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,51200,0.10890755388471816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,12288,0.04924355612860786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,128,0.0041244443919923575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,256,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,512,0.040264000495274864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,10240,0.047954665289984814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,3584,0.0262755552927653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,12288,0.015103111664454142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,512,0.2525564564598931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,512,0.02401955591307746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,512,0.22165866692860922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,3072,0.8290284474690756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,10240,0.021288888321982488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,1536,0.09817333353890313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,512,0.029464887248145208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,128,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,256,0.005028444445795483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,16384,0.026290666725900438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,7168,0.04949333270390829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,512,0.023798222343126934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,32,0.08476799726486206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,2560,0.019291554888089497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,7168,0.018274666534529794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,6144,0.04771911104520162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,3584,0.037581334511439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,3072,0.00811111099190182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,2048,0.01311288856797748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,32,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,10240,0.1306862168841892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,2048,0.028468443287743464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,1536,0.29145065943400067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,2560,0.09892444478140937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,16384,0.9999333487616645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,6144,0.01669066647688548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,1536,0.00997866690158844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,3072,0.03576177689764235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,5120,0.030936890178256567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,768,0.00815822184085846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,6144,2.9274230533175998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,1536,0.01218577805492613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,12288,0.040341334210501775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,10240,0.20511466926998564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,16384,0.02682488825586107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,128,0.0029075555503368378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,256,0.002965333354141977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,6144,0.024107555548350017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,2560,0.047863112555609814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,4096,0.01844888925552368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,3584,0.008552889029184977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,6144,0.016647110382715862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,512,0.0035137778355015647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,256,0.00908799966176351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,51200,0.10825688971413507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,10240,0.089592887295617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,64,0.08325599961810641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,3072,0.018370666437678866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,51200,0.8996399773491753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,64,0.0026986667265494666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,5120,0.010840000377760993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,7168,0.010904889139864178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,2560,0.017815111411942374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,16384,0.4116355578104655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,3072,0.006976889239417181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,1024,0.008740444150235917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,65536,0.548558235168457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,128,0.004809777769777509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,12288,20.586963229709202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,6144,0.06813689072926839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,3072,0.007993777592976889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,16384,0.028074665202034846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,51200,2.1350665622287326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,12288,0.029952890343136255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,1024,0.0058373332851462895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,65536,0.16396000650193956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,10240,0.7510533332824707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,256,0.019138667318556044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,3584,0.025294222765498694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,3584,0.02260711126857334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,1536,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,768,0.004955555415815777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,3072,0.01776622235774994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,8192,0.27116889423794216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,4096,0.023386667172114056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,51200,0.14342221948835585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,3072,0.1062506635983785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,5120,0.03172977765401205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,5120,0.017709333035680983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,2560,0.043403555949529014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,65536,0.5124435424804688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,10240,0.014073777529928418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,16384,0.048186666435665555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,65536,1.9224561055501301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,6144,0.012352888782819113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,256,0.012474666866991254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,2048,0.0998577806684706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,7168,0.021384000778198242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,6144,0.024275556206703186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,256,0.017514665921529133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,512,0.008262222011884054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,2048,0.018509333332379658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,10240,0.031153778235117596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,768,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,64,0.015019555886586508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,4096,0.027165333429972332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,32,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,256,0.004351110921965705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,65536,0.2040568855073717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,128,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,4096,0.010061333576838175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,3072,0.09401866462495591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,256,0.013409777647919126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,6144,0.15876800484127468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,7168,0.021160889003011916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,2560,0.2572711043887668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,65536,0.12137422296735977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,128,0.012886222038004132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,4096,0.1393724415037367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,4096,0.009287999735938178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,10240,0.022680888573328655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,128,0.007253333098358578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,1024,0.012896888785892062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,3072,0.04497155547142029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,3584,0.017235555582576327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,2560,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,12288,0.06198133362664116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,65536,0.2250124348534478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,1536,0.005454222361246745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,128,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,64,0.020933333370420668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,2560,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,768,0.00518844442235099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,16384,0.13761688603295222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,51200,0.2427911228603787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,2560,0.015007111761305066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,7168,0.010888000329335531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,3584,0.11967289447784424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,32,0.009463999834325578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,16384,0.49956798553466797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,64,0.01036622209681405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,8192,0.01680088871055179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,3072,0.2115244468053182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,2048,0.10032000144322713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,51200,0.6138978004455566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,256,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,512,0.007163555257850223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,1536,0.028429332706663344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,51200,0.05451200074619717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,2048,0.19749955336252847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,1536,0.012321777641773224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,16384,0.022340445054901972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,1536,0.004603555632962121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,256,0.02029511166943444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,6144,0.011379555695586734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,32,0.1546808878580729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,5120,0.043475555049048535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,256,0.007797333101431529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,1024,0.003989333493842019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,7168,0.2394017908308241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,51200,0.057541330655415855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,8192,0.16240710682339138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,5120,0.5568035443623861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,32,0.017480888300471835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,65536,0.06916977961858113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,128,0.014712888333532544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,256,0.0047493332790003884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,51200,0.025801777839660645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,65536,1.490109337700738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,3584,0.41658933957417804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,1024,0.07806399795744154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,1536,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,128,0.02675466736157735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,65536,0.2367102305094401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,3584,0.01922044489118788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,1024,0.010006222460005019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,128,0.004446222136418025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,1536,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,256,0.003540444291300244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,1024,0.0037902221083641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,3072,0.05051911208364698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,3584,0.015830222103330824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,12288,0.5082542101542155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,1024,0.007848888635635376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,5120,0.029880887932247583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,8192,0.12740177578396267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,5120,0.0064586665895250105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,2048,0.008844444321261512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,768,0.008205333517657386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,1536,0.9105670717027453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,512,0.029176000091764662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,10240,0.014847111370828418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,256,0.023892444041040208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,8192,0.016414221790101793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,64,0.002529777793420686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,2048,0.6681280136108398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,12288,0.054287999868392944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,65536,0.6036737759908041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,7168,0.0159244438012441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,1024,0.010999110837777456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,5120,0.020992888344658747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,3584,0.010677333507272931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,6144,0.014295111099878946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,51200,0.06633955240249634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,3072,0.008557333714432186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,128,0.005221333354711533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,3072,0.14976267019907633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,512,0.006726222319735422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,2560,0.01815822223822276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,7168,0.01130222198035982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,2560,0.6670764287312826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,4096,0.027083555857340496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,7168,1.8498471577962239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,32,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,65536,15.962844848632812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,16384,0.06401866674423218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,1024,0.07547999752892388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,32,0.3803306685553656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,16384,0.42393867174784344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,1536,0.026200888885392085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,16384,0.20476888285742867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,6144,1.5377271440294054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,8192,0.03176711002985636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,51200,0.016843555702103507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,16384,0.11454044448004828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,12288,0.36223822169833714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,768,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,7168,0.019671999745898776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,6144,0.012526222401195101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,16384,1.643851598103841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,4096,0.2240497801038954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,6144,0.12026666270362006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,1536,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,65536,3.6887385050455728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,512,0.022710222336981032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,256,0.004776000148720211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,3584,0.03328088919321696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,128,0.00964355551534229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,3072,0.33932889832390684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,6144,0.4212595621744792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,2048,0.0063440000845326324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,1024,0.10703911383946736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,7168,0.01757688820362091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,768,0.011511111425028907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,512,0.005530666559934616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,1536,0.03391111228201125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,1536,0.014299554957283868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,256,0.005774222314357758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,10240,0.013383111192120446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,6144,0.6177333196004232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,1536,0.1467342244254218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,5120,0.3995430999332004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,12288,0.5398613082038032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,256,0.014148443937301636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,16384,0.3367573420206706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,512,0.004982222285535601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,65536,3.1633048587375217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,3072,0.008349333372380998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,3072,0.00941777808798684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,3072,0.009763555394278632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,512,0.046317332320743136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,128,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,128,0.0038382220599386427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,32,0.006543111056089401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,1024,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,65536,0.7259217898050944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,2560,0.7540062268575033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,6144,0.13671999507480198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,1024,0.009218666288587783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,16384,0.48984267976548934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,12288,0.021849777963426378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,64,0.0069395556218094295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,512,0.027635556128289964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,1536,0.010406222608354356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,8192,0.10364355643590291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,128,0.0031022222505675424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,1024,0.018592889110247295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,3072,0.016149333781666227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,7168,0.01923733287387424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,2048,0.006226666685607698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,16384,0.052467554807662964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,768,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,8192,0.1471031109491984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,16384,0.02832977639304267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,16384,0.02509422269132402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,4096,0.010815111299355825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,16384,0.5360293388366699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,1536,0.007264888948864407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,128,0.027632888820436265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,2048,0.008621333373917473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,7168,0.10168444448047215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,8192,0.03379288978046841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,1536,0.020440000626775954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,65536,31.115648057725693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,1536,0.09713155693478054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,2560,0.011095999843544431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,1024,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,2048,0.14583910836113825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,8192,0.01200977795653873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,2560,0.08283199866612752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,128,0.0028408887899584244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,768,0.020283554991086323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,51200,0.33222844865587026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,8192,0.03673244516054789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,32,0.00508088908261723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,10240,0.006625777731339137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,6144,0.1542497740851508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,5120,0.023136888941129048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,6144,0.010204444328943888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,10240,0.02164977788925171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,6144,0.05563555823432075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,2048,0.02566755480236477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,3072,0.17140000396304658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,3072,0.007524444825119442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,10240,0.04393777913517422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,64,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,2048,0.019127999742825825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,4096,0.06438666582107544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,16384,0.14673599931928846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,4096,0.011557333171367645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,768,0.01221777747074763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,1024,0.018544889158672757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,7168,0.02711822258101569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,3072,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,256,0.005342222336265776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,3072,0.0163111107216941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,10240,0.013257777525318993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,64,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,65536,0.06796711020999484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,1536,0.006257777826653586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,5120,0.009896888501114314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,128,0.0027111110587914786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,4096,0.008207110895050896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,10240,0.2072604497273763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,256,0.0030737777964936364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,10240,0.6404017872280544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,3072,0.17251111401451957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,12288,0.23151021533542207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,51200,0.1751697725719876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,2560,0.665180418226454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,768,0.01815466748343574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,6144,0.03139644530084398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,1024,0.12793866793314615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,128,0.01109955542617374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,128,0.008115555677149031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,768,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,2048,0.04451377855406868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,7168,0.02237155536810557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,5120,0.573728879292806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,10240,0.21612355444166395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,1536,0.007370666497283512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,7168,0.01346933344999949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,3584,0.03016266557905409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,1536,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,10240,0.08065244224336413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,256,0.018111111389266122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,3584,0.25531111823187935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,6144,0.18785777356889513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,16384,0.3435795572068956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,6144,0.9288898044162326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,16384,0.07749955521689521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,128,0.006212444355090459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,7168,0.020506666766272653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,12288,0.12077689170837402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,3072,0.008018666671382057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,2048,0.11152711179521348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,4096,0.023616888456874426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,2048,0.007169778148333232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,1024,0.0038702223036024305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,12288,0.01828266680240631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,2048,0.009939555492666032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,2048,0.07284800211588542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,7168,0.017849778135617573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,10240,0.2448666625552707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,768,0.03806311223242018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,10240,0.032855110036002264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,16384,0.02279377811484867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,5120,0.016919111212094624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,7168,0.03502310978041755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,65536,0.3003244400024414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,128,0.007197333706749811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,768,0.0064488889442549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,16384,0.047685331768459745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,3072,0.019684443871180218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,3584,0.019162666466501024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,65536,0.03931022352642483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,3072,0.007070221834712558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,512,0.00867733359336853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,5120,0.039112889104419284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,64,0.004984000076850255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,51200,1.0437502331203885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,512,0.08079733451207478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,10240,0.020488889680968392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,6144,0.040030221144358315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,8192,0.8497173521253797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,16384,1.0949555502997506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,10240,0.05915021896362305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,10240,0.02515911062558492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,2560,0.032403555181291364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,10240,0.13810755146874323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,10240,0.15851110882229275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,4096,0.051262223058276706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,768,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,12288,1.1992951499091256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,5120,0.09440355830722386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,768,0.005930666708283954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,3072,0.10303022464116414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,8192,0.0322035551071167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,3072,0.008290666672918532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,7168,0.0306915541489919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,64,0.004078222231732475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,8192,0.25943732261657715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,256,0.008611555728647444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,1536,0.017850667238235474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,64,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,3072,1.3019928402370877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,65536,4.588775210910373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,256,0.013255111045307584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,32,0.003245333416594399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,16384,0.7134097947014703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,10240,0.053858667612075806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,768,0.028404444456100464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,7168,0.09340977668762207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,3072,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,8192,0.3282471232944065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,4096,0.013379555609491138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,128,0.006232888748248418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,256,0.003107555624511507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,64,0.005470222069157495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,1024,0.06934310992558797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,32,0.005469333380460739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,5120,0.012831110921170978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,6144,0.014872888724009195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,4096,0.007111999723646376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,5120,0.4186604552798801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,128,0.008092444803979661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,12288,0.045687112543318004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,1024,0.008479110896587372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,65536,0.056695110268063016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,3072,0.0555004444387224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,2048,0.03523111012246873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,128,0.004152889053026835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,16384,0.018906666172875296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,10240,0.057442665100097656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,512,0.026173333326975506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,10240,0.43363910251193577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,128,0.005076444397370021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,1536,0.01807466646035512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,6144,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,256,0.005984888722499211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,256,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,6144,0.025864000121752422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,51200,1.9187645382351344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,6144,0.03679822219742669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,64,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,3584,0.058189332485198975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,2560,0.5688071250915527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,65536,1.7438809076944988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,10240,0.06687466965781318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,16384,0.2241128815544976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,51200,0.3062257766723633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,6144,0.0571075545416938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,51200,0.049076444572872586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,768,0.0037875556283526947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,6144,0.39568177858988446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,10240,0.019666666785875957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,128,0.005049777941571342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,3584,0.016568000117937725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,65536,2.671344969007704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,32,0.009352888498041365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,768,0.1419022215737237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,51200,3.0267921023898654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,5120,0.09302844603856404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,8192,0.031077331966824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,16384,0.06360799736446805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,512,0.009184888667530483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,51200,1.6469270918104384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,12288,0.030925333499908447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,8192,0.01924088928434584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,12288,0.04240622123082479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,1024,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,6144,0.6539973153008355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,3072,0.018056000272432964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,65536,0.19116444057888457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,768,0.011035555766688453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,128,0.004423999951945411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,12288,0.007206222249401941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,2560,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,1024,0.004051555362012652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,4096,0.01846844454606374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,512,0.026919111609458923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,32,0.012418666647540199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,64,0.034494221210479736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,256,0.004576888763242298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,10240,0.02310755517747667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,7168,0.011930666863918304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,6144,0.30948710441589355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,4096,2.274855931599935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,3584,0.009146666361225976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,3584,0.016799110505315993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,8192,0.15831111537085638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,1536,0.006134222365087933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,51200,0.05517333414819506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,12288,0.009033777647548253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,768,0.016413332687483895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,4096,0.03427111109097799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,4096,0.3207288848029243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,128,0.04508089025815328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,65536,0.32587112320794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,768,0.004591111093759537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,4096,0.1334542168511285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,12288,1.8887030283610027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,16384,0.30275734265645343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,512,0.0034062222888072333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,1536,0.061090668042500816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,6144,0.013378666506873237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,128,0.003232888877391815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,3072,0.04705333378579882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,256,0.005135111096832487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,64,0.006592888798978593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,2560,0.008314666648705801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,128,0.0032426667296224167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,51200,0.10549955235587226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,10240,0.022061333060264587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,128,0.005474666754404704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,768,0.008239111138714684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,6144,0.09316444396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,3584,0.22946577601962617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,2048,0.026018665896521673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,8192,0.0728400018480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,16384,0.06427822510401408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,8192,0.1897102197011312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,51200,0.4218417803446452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,51200,3.6947716606987844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,10240,0.024324445260895625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,7168,0.021155556042989094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,1536,0.0064177778032090925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,16384,0.03495733274353875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,16384,0.035094221433003746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,512,0.01891999940077464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,10240,0.5172995461357964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,12288,0.04073422153790792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,12288,0.022315555148654517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,128,0.0036720000207424164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,51200,0.2200382285647922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,65536,0.0652408864763048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,512,0.01961777773168352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,32,0.009934222532643212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,7168,0.07847022347980075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,3072,0.02814488940768772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,7168,0.009525333013799455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,4096,0.017712000343534682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,16384,0.029704888661702473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,2560,0.013467555244763693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,3584,0.009947555760542551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,32,0.0026088888860411118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,128,0.007640889121426477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,32,0.0029057777590221832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,3072,0.07320088810390897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,64,0.009621333744790819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,12288,0.08918399943245782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,8192,0.1969857745700412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,3072,0.010107555323176915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,65536,1.007923550075955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,51200,0.18121777640448677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,256,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,8192,0.013759999639458127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,128,0.003456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,2560,0.0075333333677715724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,256,0.009692444569534725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,32,0.0046693334976832075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,2560,0.007170666423108842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,12288,1.4975182215372722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,2048,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,768,0.2784586747487386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,768,0.02515911062558492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,12288,0.12596978081597224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,256,0.04514844550026787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,51200,0.06275644567277697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,2048,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,8192,0.015232000086042615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,2048,0.04118844535615709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,2560,0.05040711164474487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,512,0.027795556518766616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,768,0.017119111286269296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,65536,0.8892186482747396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,8192,0.011764444410800934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,16384,1.0100328657362196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,64,0.0736959973971049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,1536,0.12618045012156168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,65536,0.2175511121749878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,512,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,3072,0.041624890433417425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,65536,0.015257777439223396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,6144,0.022328888376553852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,65536,0.5181368721856011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,64,0.005229333208666907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,2560,0.19970044824812147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,3584,0.01885333326127794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,768,0.006134222365087933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,2048,0.04012711180580987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,8192,6.572414398193359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,65536,0.08523199955622356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,6144,0.12046133147345649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,128,0.02409155501259698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,12288,0.034691555632485285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,8192,0.05728089147143894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,4096,0.030736890104081895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,1024,0.05131999982727898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,128,0.004001777619123459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,2560,0.12279021739959717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,2560,0.0173680004146364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,4096,0.19485155741373697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,12288,0.3289262188805474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,2048,0.007095110913117726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,8192,0.024177778098318312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,32,0.06829955842759874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,1024,0.28617776764763725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,128,0.0046880000995265115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,128,0.021760000122918025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,51200,51.944186740451386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,256,0.003913777983850903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,7168,0.030770666069454614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,1536,5.163823869493273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,2048,0.015586665934986539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,768,0.006701333241330252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,32,0.006771555377377405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,64,0.005250666704442766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,3584,0.25421423382229275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,2560,0.008411555654472774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,3584,0.17572000291612414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,51200,0.9696764416164823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,1024,0.008789333204428354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,4096,0.13314399454328749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,2560,0.25110843446519643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,3072,0.01759199963675605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,3072,0.11457599533928765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,512,0.05037777622540792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,5120,0.011343999869293638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,4096,0.019659555620617337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,3072,0.011117333339320289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,10240,0.34046043290032285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,8192,0.023171555664804246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,128,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,256,0.006151999864313338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,128,0.0029502221279674103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,65536,2.087096956041124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,16384,0.41286489698621964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,16384,0.16169599692026773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,2048,0.01566844516330295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,2048,0.01707466608948178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,4096,0.009119110802809397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,512,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,256,0.00888355573018392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,7168,0.013191999660597907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,256,0.00443466669983334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,7168,0.03103644318050808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,8192,0.547203540802002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,65536,4.977624681260851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,7168,0.01591111057334476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,256,0.00423200014564726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,65536,7.121842278374566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,768,0.025054223007626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,4096,0.009032888544930352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,51200,0.9812471601698133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,1536,0.0096124443742964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,6144,0.01570399933391147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,4096,0.05738044447369046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,128,0.007787555456161499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,64,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,12288,1.4363270865546331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,1536,0.24274844593471953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,51200,0.45741690529717344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,128,0.0032515554792351196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,4096,0.1289315621058146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,12288,0.1328737735748291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,64,0.024521778027216595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,512,0.011808888779746162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,6144,0.06767466995451185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,6144,0.08160799741744995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,4096,0.006953777538405524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,4096,0.016295111841625638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,16384,0.018591110905011494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,51200,0.19058222240871855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,12288,0.05233777893914116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,65536,0.010064000056849586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,4096,0.07544177770614624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,7168,0.020082667469978333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,8192,0.3693866729736328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,1024,0.051832000414530434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,51200,0.07969511217541166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,16384,0.04191822144720289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,1024,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,51200,0.6798728836907281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,64,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,3584,0.47527111901177305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,1536,0.06015111340416802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,16384,0.1391164461771647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,7168,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,3072,0.015457777513398064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,32,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,3072,0.009172444542249044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,65536,0.15013689464992946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,7168,0.09931021928787231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,6144,0.012114666402339935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,512,0.004903111192915174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,5120,0.013891556196742587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,128,0.003630222131808599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,3584,0.11076800028483073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,10240,0.015923556354310777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,2048,0.02474755545457204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,3072,0.009988444546858469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,65536,1.7903351253933375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,16384,0.2788444360097249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,128,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,12288,0.02068000038464864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,65536,2.3182782067192926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,2048,0.017462222112549674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,256,0.016896888613700867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,8192,0.21306133270263672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,7168,2.6349991692437067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,12288,0.01499466598033905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,6144,0.015012444721327888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,1536,0.015537778536478678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,1536,0.007621333003044128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,2048,0.02032800018787384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,1536,0.2638817893134223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,256,0.12497511174943711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,2560,0.014493332968817817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,7168,0.3610275586446126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,1536,0.005254222287072076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,128,0.003528000166018804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,1024,0.010081777969996134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,64,0.005698666804366642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,8192,0.09526311026679145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,128,0.004661333229806689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,1536,0.007694222033023834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,2048,0.011959111524952782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,16384,0.20112178060743544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,51200,0.5389439794752333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,128,0.05778488847944471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,12288,0.04721066686842176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,5120,0.010050666828950247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,8192,0.03851911094453599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,16384,0.5679617987738715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,5120,0.012817777693271637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,128,0.02812266680929396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,32,0.006514666808976068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,3584,0.008609777523411645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,6144,0.06807021962271796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,65536,0.2586870988210042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,65536,0.1115493377049764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,12288,0.04535288943184746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,4096,0.012551111479600271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,2048,0.007605333295133378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,256,0.006148444281684027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,2560,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,256,0.006917333437336816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,3584,0.008481777376598781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,1536,0.05122844378153483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,768,0.012028444144460889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,256,0.004676444249020683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,10240,0.19540533754560682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,128,0.004109333372778363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,8192,0.31785421901279026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,16384,0.5395449002583822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,10240,0.11762577957577175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,51200,0.15462399853600398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,512,0.003351111171974076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,10240,1.9354657067192926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,8192,0.07265955872005887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,16384,0.02169599963559045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,5120,0.06450577576955159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,128,0.004057777838574516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,10240,0.01697422232892778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,32,0.005414222263627582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,512,0.03404622276624044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,7168,0.9757039811876086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,3072,0.010553778045707278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,51200,2.901639090643989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,7168,0.01531022290388743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,3584,0.19116711616516113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,6144,0.008240000241332585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,10240,0.03142222099834018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,256,0.02383111086156633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,7168,0.01292799992693795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,256,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,3584,0.0184906671444575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,64,0.0028008888992998335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,3072,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,1024,0.005766222046481238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,65536,1.191056039598253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,3584,0.03604088889227973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,4096,0.3032986587948269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,8192,0.0728400018480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,3584,0.013088000317414602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,1536,0.5102737744649252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,128,0.004392000122202767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,128,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,768,0.011035555766688453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,3584,0.006375999914275275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,128,0.0669742226600647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,1024,0.25987111197577584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,65536,0.16045244534810385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,64,0.0027173333283927706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,5120,0.014177777700954013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,128,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,16384,5.619461483425564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,512,0.007182222273614671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,128,0.005655111124118169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,65536,0.07398577531178792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,8192,0.045780443482928805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,512,0.027458666099442378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,768,0.01404266721672482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,12288,0.00756444450881746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,512,0.005852444304360284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,5120,0.01703555550840166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,16384,0.15790488984849718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,768,0.006984000404675801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,32,0.0052782222628593445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,256,0.008073777788215214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,51200,0.16824266645643446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,3072,0.14058044221666124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,1024,0.06365778048833211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,4096,0.030454221698972914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,768,0.17006044917636445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,1536,0.07011733452479045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,8192,0.1061955557929145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,768,0.02251288957066006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,768,0.15591555171542698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,6144,0.19545155101352266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,5120,0.927512910630968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,51200,0.7124773131476508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,2560,0.01644888851377699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,1024,0.005769777629110549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,1024,0.008044444852405125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,65536,1.2850009070502388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,128,0.036222222778532244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,1536,0.059593776861826576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,10240,0.34062578943040633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,4096,0.010644444160991244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,4096,0.013539555172125498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,5120,0.32567911677890354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,6144,0.0207324441936281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,4096,0.03308355477121141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,5120,0.031081779135598078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,1536,0.013622221847375235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,32,0.007530666887760162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,256,0.005584888988071018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,32,0.007888000044557784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,2560,0.015213333898120456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,128,0.031890667147106595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,8192,0.06611733304129706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,256,0.008831111093362173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,3584,0.008984000318580205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,512,0.09476266966925727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,8192,0.3359946674770779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,2048,0.01942933268017239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,16384,6.162955390082465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,3072,0.17621866861979166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,10240,0.4949626922607422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,64,0.015481778317027621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,6144,0.010975110861990186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,512,0.004119111018048393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,1536,0.008984888593355814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,768,0.008563555777072906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,3072,0.009687999884287516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,5120,0.0344817770851983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,512,0.059691554970211454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,7168,0.02346044447686937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,16384,0.1675635576248169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,51200,0.6938240263197156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,16384,0.0884595579571194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,2048,0.03903733359442817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,2560,0.0159288893143336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,32,0.02845066785812378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,256,0.02510133385658264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,3072,0.1055262221230401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,10240,0.03564800156487359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,512,0.005790222022268508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,5120,0.03648444347911411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,256,0.0029804443733559716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,256,0.003023111157947116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,1024,0.013562666873137156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,3584,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,768,0.004263111286693149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,3584,0.02347466680738661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,2048,0.016418667303191293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,7168,0.12497333685557048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,2560,0.01738755570517646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,2048,0.010808889236715106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,8192,3.356675465901693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,7168,0.04699555701679654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,2048,0.1285368866390652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,5120,0.015800888339678448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,6144,0.025049777494536504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,6144,0.010952889091438718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,3584,0.31454043918185765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,256,0.05415466758939955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,32,0.012501333322789935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,5120,0.04994844396909078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,512,0.0034026667061779234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,7168,0.12813333670298258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,3072,0.007528000407748752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,768,0.003573333223660787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,16384,0.04312977857059903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,64,0.002471111093958219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,10240,0.013470222552617392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,7168,0.013285333083735572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,51200,0.0693146652645535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,2560,0.10997511280907525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,128,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,64,0.2691475550333659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,51200,0.14110310872395834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,768,0.009798222117953831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,3584,0.011448889142937131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,4096,0.9091261757744683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,32,0.10439378023147583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,16384,0.01611911091539595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,5120,0.08862844440672134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,3072,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,7168,0.028007109959920246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,32,0.006070222291681502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,12288,0.031231999397277832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,2048,0.07666933536529541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,128,0.0039013334446483185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,12288,0.02516799999607934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,64,0.0031439999325407874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,512,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,512,0.003742222156789568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,6144,0.028584887584050495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,51200,1.0647226969401042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,16384,0.07220977544784546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,2048,0.004977777600288391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,1024,0.007495111061467065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,512,0.011346666349305047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,51200,0.04978311061859131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,512,0.003653333418899112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,128,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,768,0.007829333345095316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,1536,0.7722044520907931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,16384,0.05448977814780342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,32,0.02325244413481818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,12288,0.021127111381954614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,51200,0.07823199696011014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,7168,0.8744515313042535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,128,0.02370133333735996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,51200,0.14125600126054552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,1024,0.01418133411142561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,1536,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,512,0.10792977942360772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,128,0.006007110906971826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,64,0.0028853333658642242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,256,0.013468444347381592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,6144,0.012319111161761813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,10240,0.19479289319780138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,4096,0.012168889244397482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,3072,0.012119111087587146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,12288,0.4448551072014703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,4096,2.3046692742241754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,10240,0.025935999221271936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,4096,0.012612444659074148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,4096,0.013451555536852943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,768,0.03640888796912299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,2560,0.05265511075655619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,128,0.031665778822369046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,3072,0.010075555907355415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,1024,0.006411555740568373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,5120,0.16132177246941462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,5120,0.05549600058131748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,512,0.0046977777447965415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,256,0.004783111313978831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,6144,0.03713599840799967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,12288,0.5967671076456705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,128,0.007268444531493717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,10240,0.03168088859981961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,128,0.021663111117151048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,3584,0.009886222581068674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,10240,0.050829334391487964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,1536,0.026914666096369427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,768,0.005952000204059813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,512,0.004490666505363253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,8192,0.2694142129686144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,12288,0.04632088873121473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,4096,0.010495111346244812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,5120,0.20562399758232963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,51200,0.14198488659328884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,6144,0.07495733102162679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,10240,2.179410722520616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,256,0.010455999937322406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,6144,0.014159111513031853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,768,0.05139466789033678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,12288,0.49240711000230575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,65536,0.08841244379679362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,3584,0.023224888576401606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,128,0.003506666670242945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,5120,0.10804710785547893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,65536,0.07713511255052355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,65536,1.0520835452609592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,1024,0.006692444284756978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,256,0.006611555400821898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,7168,0.21815110577477348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,2560,0.02788177794880337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,1024,0.020763556162516277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,2560,0.044512000348832875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,3072,0.014879110786649915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,12288,0.18221778339809844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,2048,0.05704622136221992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,32,0.008286221987671321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,128,0.010589333044158088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,4096,0.013445333474212222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,4096,0.07540444533030193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,768,0.0991484456592136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,32,0.010805333654085795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,3584,0.09569244252310859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,16384,0.05460266603363884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,64,0.008238222036096785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,10240,0.05997688902748955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,16384,0.3014337751600477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,3584,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,10240,0.05894933144251505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,3584,0.024777778320842322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,64,0.00814577771557702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,512,0.004515555583768421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,6144,0.01091733326514562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,2048,0.011448889142937131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,32,0.028881778319676716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,2048,0.04905688762664795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,16384,6.401941511366103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,8192,0.026105778084860906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,65536,0.10043111112382676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,12288,0.1860426664352417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,3072,0.06339289082421197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,768,0.004338666796684265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,64,0.0057839999596277876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,16384,0.04907288816240099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,1536,0.01459022197458479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,3584,0.011744889120260874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,32,0.05670311053593954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,10240,0.3274808989630805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,8192,0.01168444421556261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,10240,0.03012711140844557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,32,0.019944889677895438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,512,0.005433777968088786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,6144,0.008239111138714684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,1536,0.01385600037044949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,10240,0.018174222773975797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,3072,0.024362666739357844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,8192,0.026837334036827087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,4096,0.04197600152757433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,256,0.0038399998512532976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,3584,0.012152888708644442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,5120,0.012294222083356647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,8192,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,5120,0.5572711096869575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,32,0.0053342220683892565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,2048,0.20774843957689074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,1536,0.025753777888086107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,10240,0.416098674138387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,16384,1.1756347020467122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,3584,0.01945600079165565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,256,0.010470222267839642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,2560,0.012984889249006907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,768,0.05498577819930183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,2560,0.13080444600847033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,10240,0.018031999468803406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,32,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,7168,0.24517867300245497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,6144,0.11065066523022121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,512,0.006884444504976273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,64,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,10240,4.809382968478732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,8192,0.24061332808600533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,8192,0.11898933516608344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,4096,0.019638222124841478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,6144,0.008575999902354347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,2048,0.13188710477617052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,5120,0.006519111080302133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,2560,0.006088888893524806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,7168,0.022318222456508215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,51200,0.06887822018729316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,16384,0.023579556081030104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,64,0.0026488889836602737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,6144,0.02017955482006073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,1536,0.016833777228991192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,512,0.08949689070383708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,1024,0.037131554550594754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,3072,0.009191110730171204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,65536,0.1066711081398858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,7168,0.15167289310031468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,64,0.007038222418891058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,5120,0.1174817747539944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,3072,0.03433600068092346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,7168,0.0061039999127388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,128,0.004908444566859139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,5120,0.04145155681504144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,16384,0.17061955398983428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,5120,0.016786666380034555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,32,0.008184000021881526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,6144,0.46712530983818906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,5120,0.03611733184920417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,12288,0.12058399783240424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,2048,0.006509333435032103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,1536,0.02051911089155409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,10240,0.025343111819691126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,51200,0.145578662554423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,768,0.008042666647169325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,6144,0.2545599937438965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,1536,0.10438755485746597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,6144,0.041857779026031494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,256,0.04437511165936788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,4096,0.05213155680232578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,6144,0.012351999680201212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,512,0.006142222219043308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,4096,0.10425333182017009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,51200,0.6642657915751139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,2560,0.46361154980129665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,64,0.010799999866220685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,128,0.005038222091065513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,16384,0.2509555551740858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,4096,0.019154666198624503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,256,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,16384,1.2646968629625108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,256,0.0052737775776121355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,12288,0.3687617778778076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,7168,0.025076443950335186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,12288,0.23067288928561738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,4096,0.01995377739270528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,65536,4.547935909695096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,512,0.03562577896647983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,1024,0.007527111305130853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,3072,0.22316710154215494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,3584,0.011797332929240333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,3584,0.49126222398546004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,256,0.01575999955336253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,1024,0.0068604445291890045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,65536,0.08308177524142794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,4096,0.03158311049143473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,8192,0.08331911431418525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,512,0.015625778171751235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,3584,2.806264029608833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,256,0.02455377744303809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,7168,0.06073421902126736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,1024,0.007172444628344641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,3072,0.007837333612971837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,2560,0.06260710954666138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,8192,0.046259555551740855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,65536,0.09012622303432888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,512,0.009619555539555019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,6144,0.012337778177526263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,10240,0.01777688827779558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,7168,0.02160888910293579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,65536,0.7198950979444715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,128,0.07902044720119901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,1536,0.05835555659400093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,12288,0.33470932642618817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,4096,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,128,0.006844444407357111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,256,0.0031822222388452957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,65536,0.30026311344570583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,3072,0.01105955574247572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,1024,0.004936000125275718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,1024,0.004101333518822988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,3584,0.01054933336046007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,2048,0.07392177979151408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,256,0.01016444464524587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,2048,0.005375111268626318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,1536,0.004644444419278039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,512,0.017753778232468497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,3072,2.0792959001329208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,16384,1.2564364539252388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,5120,0.13558933469984266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,65536,1.3904790878295898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,3584,0.012489777472284106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,65536,0.5073919826083714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,2048,0.01631733278433482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,10240,0.12691111034817165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,4096,0.013431111143694984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,32,0.034091555409961276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,12288,0.0703751113679674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,3072,0.00905600024594201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,7168,0.013337777720557319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,4096,0.0064026667839950984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,128,0.004505777938498391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,1024,0.046774221791161426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,768,0.005682666682534748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,1536,0.010915555887752108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,4096,0.013382222089502545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,1024,0.012809777425395118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,12288,0.10596177975336711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,32,0.004451555510361989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,512,0.02664711078008016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,256,0.02388977838887109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,1536,0.006445333361625671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,128,0.0027635554886526535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,3584,0.11191288630167644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,2560,0.2506648964352078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,2560,0.011959111524952782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,12288,0.017998221847746108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,1536,0.011610666910807291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,256,0.004981333182917701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,12288,0.1409075525071886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,2048,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,128,0.0026497778793176017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,768,0.09340355793635051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,32,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,1024,0.08933866686291164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,2560,0.054791112740834556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,512,0.00452533322903845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,65536,54.75055609809028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,51200,2.868918312920464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,51200,0.14799733956654867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,10240,0.01947733263174693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,4096,0.10270844565497504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,12288,8.688139173719618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,65536,0.5476213561164008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,768,0.008250666989220513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,64,0.007883555359310573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,16384,1.8056774139404297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,4096,0.1592506700091892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,7168,0.01479022204875946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,1024,0.006671999891599019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,1536,0.027588443623648748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,2560,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,7168,0.015452444553375244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,256,0.0034764446318149567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,128,0.017826666434605915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,12288,0.015535111228624979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,1024,0.010840888652536603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,32,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,3072,0.1210133367114597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,768,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,2048,0.01585155559910668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,3584,0.01872977779971229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,512,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,1536,0.05004622207747566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,1536,0.0262071109480328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,8192,0.048786666658189565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,65536,0.17436622248755562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,3584,0.06877866718504164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,5120,0.012327999704413943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,128,0.004782222211360931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,8192,0.011747555600272285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,4096,0.22360621558295357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,256,0.019766221443812054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,65536,0.06608977582719591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,1024,0.051623112625545926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,8192,0.17439644866519502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,1536,0.00905600024594201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,5120,0.0063075555695427795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,256,0.007710222568776872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,2048,0.016373333003785875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,1024,0.025751999682850305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,7168,0.07840089003245036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,768,0.008868444297048781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,10240,0.07575911283493042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,3584,0.011776000261306763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,16384,0.23207555876837838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,65536,0.1420417759153578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,3072,0.03323466579119364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,512,0.10772889190249973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,3584,0.03456000155872769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,32,0.005142222262091107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,256,0.008259555531872643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,2048,0.16435466872321233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,512,0.011410666836632622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,1024,0.010121777653694153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,768,0.004967999955018361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,16384,0.37055643399556476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,128,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,10240,0.07522755861282349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,8192,0.014442667365074158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,2560,0.08018222120073107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,10240,0.9397528966267904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,7168,0.014656000667148165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,768,0.007366221812036302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,1536,0.1148444414138794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,64,0.0047848886913723415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,256,0.0031031111462248694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,8192,0.02314044369591607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,5120,0.1781733300950792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,16384,1.5342186821831598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,256,0.005533333453867171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,8192,0.04407377706633674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,4096,0.024634665913052026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,2048,0.10450933376948039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,512,0.0063742221229606206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,4096,0.06658577919006348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,128,0.004267555558019214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,5120,0.07556177510155572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,256,0.005573333137565189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,2048,0.010847999817795224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,512,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,7168,0.5507235527038574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,512,0.035487999518712364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,768,0.005602666487296422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,5120,0.036299556493759155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,4096,0.012216000093354119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,8192,0.015677332878112793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,3072,0.26358310381571454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,65536,0.07080711258782281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,12288,0.02164711058139801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,5120,0.29265954759385854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,51200,0.07478844457202487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,64,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,768,0.005362666729423735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,7168,0.014304000470373364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,65536,1.2314702139960396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,1536,0.007178666690985362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,64,0.050070222881105214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,5120,0.019332443674405415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,5120,1.4145431518554688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,2048,0.015242666006088257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,16384,0.07827022340562609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,128,0.006710222197903528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,1536,0.007984889050324758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,16384,0.06074133184221056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,8192,0.22269156244066027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,128,0.0028080000645584534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,2048,0.006731555693679386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,768,0.15468622578514948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,10240,0.04770400126775106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,32,0.0042133331298828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,51200,0.7371893458896213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,32,0.002945777856641345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,6144,0.032483554548687406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,2560,0.07870933082368639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,1024,0.00647288892004225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,16384,1.6102693345811632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,2048,0.015829333000712924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,16384,2.0420896742078996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,768,0.009322666459613377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,5120,0.01761244402991401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,10240,0.009397333694828881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,4096,4.222482469346788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,2048,0.0075253330998950535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,3584,0.44642845789591473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,512,0.006801777829726537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,16384,0.08065955506430732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,3072,0.1583848926756117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,2560,0.013253333667914072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,16384,0.021726222501860723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,65536,0.25407910346984863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,32,0.01423822177780999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,2560,0.009722666607962714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,32,0.0026764445420768526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,16384,0.28786222139994305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,4096,0.026206221845414903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,65536,0.08869244654973348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,64,0.002789333255754577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,7168,0.9442906909518771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,3072,0.744790236155192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,256,0.06714310910966662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,1024,0.004662222332424588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,51200,0.08529333273569743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,4096,0.2832835515340169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,16384,0.04151022103097703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,32,0.004218666503826777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,768,0.004035555654101902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,512,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,5120,0.477584891849094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,2560,0.11800177892049153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,512,0.005383111122581694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,8192,2.0338683658175998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,51200,0.07223911417855157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,32,0.006760889043410619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,4096,0.014001778430408902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,128,0.0037173334923055437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,8192,0.018399111098713346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,3584,0.0306275553173489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,3072,0.015570667054918079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,2048,0.012271111210187277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,1536,0.004976888911591636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,12288,0.01924000018172794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,512,0.0034311111602518293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,128,0.023715555667877197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,51200,0.7174693213568794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,2560,0.022886221607526142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,3584,0.026388444834285315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,2048,0.005632888939645555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,4096,0.1339093314276801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,10240,0.0310444434483846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,6144,0.010829333629873065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,4096,0.009504000345865885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,1536,0.015786666009161208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,256,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,12288,0.01536355581548479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,7168,0.05482222305403816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,5120,0.018221333622932434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,4096,0.041598220666249595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,10240,0.006534222099516127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,5120,0.009399111072222391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,3584,18.79121568467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,12288,1.2651875813802083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,16384,0.026064000195927087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,6144,0.07549066676033868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,7168,47.086042616102425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,8192,0.05281955666012234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,8192,0.6945110956827799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,2048,0.03702488872740004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,2560,0.13675467173258463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,32,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,1024,0.005211555709441503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,12288,0.41174133618672687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,2560,0.1448462274339464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,6144,0.21674133671654594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,768,0.030584888325797185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,256,0.007299555672539606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,4096,0.011967110964987012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,32,0.00434400017062823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,10240,0.7639075385199653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,10240,0.011798222031858234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,65536,6.584007263183594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,51200,0.06930222113927205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,2560,0.018422222799725003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,4096,0.03471466567781236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,64,0.0025066667132907445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,3584,0.08439733584721883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,5120,1.5026497311062281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,2048,0.019088000059127808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,6144,0.017110221915774874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,3072,0.041159109936820135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,3584,0.2764071093665229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,3072,0.010939555863539377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,2560,0.01277599980433782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,512,0.003664888855483797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,10240,0.20081510808732775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,10240,0.284717321395874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,5120,0.03435644507408142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,768,0.005988444305128521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,4096,0.01939377850956387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,4096,0.01367200000418557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,1024,0.0063902222447925145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,5120,0.7178959846496582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,512,0.004254222330119875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,7168,0.013414222333166333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,65536,0.49715020921495223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,3584,0.008628444539176093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,64,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,5120,0.040766222609413996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,2560,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,10240,0.5168595314025879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,128,0.008199999729792276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,16384,0.0514373348818885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,16384,0.10876355568567912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,12288,0.3039013279808892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,256,0.008535111116038429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,512,0.004660444541109933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,64,0.002453333387772242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,256,0.06091199980841743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,2560,0.008764444126023186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,512,0.0033653332955307434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,3072,0.021697777840826247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,7168,0.03793155484729343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,512,0.005573333137565189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,8192,0.09323022100660537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,3072,0.02685333291689555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,128,0.011710222396585675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,3584,0.008809777597586313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,1024,0.003867555409669876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,2560,0.011157333023018308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,64,0.0034453332838084963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,8192,0.285624901453654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,16384,0.030192000998391047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,7168,0.0251217782497406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,65536,0.49708710776435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,64,0.22991734080844453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,256,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,4096,2.9863048129611545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,64,0.0034355556385384668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,6144,0.01672266589270698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,8192,0.1475413375430637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,256,0.01349066694577535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,8192,0.013970666461520724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,32,0.013712888790501488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,2560,0.0070159998204973005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,12288,0.6760017606947156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,8192,0.027440889014138117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,512,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,6144,0.016774222254753113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,768,0.3624346786075168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,6144,0.02382933265633053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,64,0.0033768889390759995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,2560,0.01388177772363027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,3072,0.122524446911282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,6144,0.047036445803112455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,10240,0.6027128961351182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,64,0.007720888488822513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,768,0.08969777822494507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,32,0.006728000111050076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,128,0.0033271111961868075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,256,0.03853333327505324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,3584,0.023936000135209825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,3584,0.008573333422342936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,2560,0.00813333359029558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,7168,0.044785777727762856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,256,0.0037537775933742523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,64,0.006568000134494569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,51200,0.41593689388699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,512,0.005391110976537068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,3072,0.03704266746838888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,1536,0.007621333003044128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,2048,0.012151999606026543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,3072,0.5660079850090874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,4096,0.0162791113058726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,5120,0.016840888394249808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,32,0.06026133563783434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,32,0.008634666601816813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,10240,0.043771555026372276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,7168,2.131793763902452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,768,0.026054221722814772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,1024,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,16384,0.027513777216275532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,8192,0.04249599907133314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,3584,0.009962666365835402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,12288,0.041781332757737905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,3584,0.011861333416567909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,16384,0.3140328990088569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,2048,0.08158844709396362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,1536,0.004832889056868023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,768,0.04040088918473985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,3072,0.26528088251749676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,512,0.003630222131808599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,7168,0.05208977725770739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,768,0.15077866448296443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,3584,0.015606222881211175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,64,0.007667555577225155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,32,0.014496889379289416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,1024,0.008854221966531541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,4096,0.0331928895579444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,3584,0.016763556334707472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,12288,0.046012444628609546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,3072,0.025009777810838487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,256,0.007822222179836696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,3072,0.015511999527613321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,5120,0.49878491295708555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,3072,0.021789333886570398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,128,0.006383111079533895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,8192,0.03190933333502875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,2560,0.006031111296680238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,16384,0.018477333916558158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,7168,0.020613332589467365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,1536,0.006275555739800136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,7168,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,16384,0.03891911109288534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,1536,0.006161777923504512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,2560,0.04312622216012743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,2560,0.06455200248294406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,3072,0.10530400276184082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,3584,0.00980088859796524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,7168,0.018311111463440787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,3072,0.01107022249036365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,10240,0.15714222855038112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,1536,0.006161777923504512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,6144,0.03586133321126302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,512,0.007243555453088548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,7168,0.01293155550956726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,2048,0.030375109778510198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,65536,2.472419526841905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,1536,0.06527911292182074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,5120,0.12694844934675428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,3584,0.6931040022108289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,1024,0.03168355425198873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,3072,0.009701333112186855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,3584,0.0812693304485745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,3584,0.014436443646748861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,3072,0.48721864488389754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,512,0.01829333272245195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,3072,0.009974222216341231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,5120,0.02678044471475813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,3072,0.013510222236315409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,512,0.015097777048746744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,256,0.0029973333908451926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,6144,0.010431999961535135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,4096,0.008531555533409119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,768,0.011964444484975604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,768,0.03813866774241129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,256,0.004559999952713649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,2560,0.03956800036960178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,128,0.011992000043392181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,128,0.004451555510361989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,3072,0.1531964408026801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,64,0.3393208980560303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,32,0.006710222197903528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,1024,0.003960000144110786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,256,0.005371555685997009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,65536,0.40290400716993546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,7168,0.01371911085314221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,512,0.005631999837027655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,12288,0.049286223120159574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,32,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,128,0.007132444116804335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,768,0.0075635554061995606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,512,0.01703644461101956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,1536,0.5192755593193902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,6144,0.30973243713378906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,1536,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,1024,0.006248000181383557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,512,0.005013333426581489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,12288,0.20451288753085664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,512,0.007169778148333232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,4096,0.02125511070092519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,2048,0.010102222363154093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,128,0.010845333337783813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,6144,0.020502222908867728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,1536,0.004861333303981357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,65536,0.10698133044772679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,1024,0.007516444557242923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,16384,0.03252799974547492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,256,0.004889777965015835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,12288,3.171112060546875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,128,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,1024,0.018837332725524902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,128,0.003549333247873518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,1536,0.006199999815887875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,128,0.003590222034189436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,8192,0.021871111459202234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,2560,0.09429777993096246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,64,0.0038159998754660287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,2560,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,12288,0.024920889072948035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,8192,0.019725332657496136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,6144,0.01907466683122847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,8192,0.14177066749996609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,5120,0.017614222235149808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,10240,0.07002844413121541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,8192,0.15497067239549425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,32,0.002732444347606765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,8192,0.3464133209652371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,1536,0.013944889108339945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,512,0.005013333426581489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,6144,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,768,0.004685333205593957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,1536,0.022011554903454248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,16384,0.3664711051517063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,64,0.007118222614129384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,51200,1.3965910805596247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,1536,0.007440889047251807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,64,0.1491031116909451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,32,0.002966222249799304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,512,0.005422222117582957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,32,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,256,0.011223110887739392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,128,0.006232888748248418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,7168,0.012004444168673621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,6144,0.00831288927131229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,6144,0.5038560231526693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,64,0.013420444395807056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,5120,0.023714666565259297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,65536,0.17777332994672987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,4096,0.08008088668187459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,65536,0.09136711226569282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,10240,0.038479109605153404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,8192,0.35335644086201984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,3584,0.01936444474591149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,3584,0.15375022093454996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,256,0.014069333672523499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,256,0.003528888854715559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,1536,0.006452444526884291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,3584,0.009312888814343346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,7168,0.4919564459058974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,8192,0.021783999270863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,256,0.004540444248252445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,768,0.017825777331988018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,12288,0.05512533254093594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,768,0.004982222285535601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,4096,0.016216889023780823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,10240,0.057433777385287814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,10240,0.48018312454223633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,10240,0.02923733327123854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,3584,0.03038311004638672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,2560,0.10352089007695515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,12288,0.022801778382725183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,1536,0.12131467130449082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,8192,0.025851555996470984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,5120,0.016540444559521146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,10240,0.009855999714798397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,51200,0.19273777802785239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,65536,0.5287200080023872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,32,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,1536,0.0074311114019817775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,12288,0.04564177658822802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,12288,0.06551200151443481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,2048,0.005936000082227919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,5120,0.00924177798959944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,2048,0.009400889277458191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,5120,0.02409511142306858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,51200,0.10927289062076145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,6144,0.012854222622182635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,128,0.04958666695488823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,1024,0.1573902236090766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,3072,0.02035911050107744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,6144,0.15970577134026423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,3584,0.013956444131003486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,65536,0.36778489748636883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,3072,0.007992000215583378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,3584,0.06786933210160997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,256,0.017817777064111497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,51200,7.741696675618489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,7168,0.1276035573747423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,6144,0.019000000423855253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,256,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,8192,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,65536,22.290632459852432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,64,0.012098666694429187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,3072,0.010346666806273991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,64,0.16483555899726018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,256,0.006811555474996567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,51200,0.8635582394070096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,3072,0.014582221706708273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,3584,0.0527724458111657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,2048,0.09508088562223647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,256,0.02212711175282796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,10240,0.658174196879069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,768,0.007747555772463481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,768,0.016984888248973422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,256,0.06295200188954671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,1024,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,64,0.003040888864133093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,3584,0.009669333696365356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,7168,0.039613333013322614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,512,0.005184000151024925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,8192,0.15609599484337702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,256,0.003117333269781537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,51200,0.13620089160071477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,4096,0.4435217645433214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,3584,0.033391998873816595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,2048,0.03028800090154012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,51200,0.3644399907853868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,5120,0.058246221807267934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,1024,0.009127111070685917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,16384,0.023393778337372675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,7168,0.020894222789340552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,3072,0.05438933438724942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,128,0.023702222439977858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,1536,0.08666044473648071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,3584,0.008436444732877944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,4096,0.2216728925704956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,12288,0.046075556013319224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,1024,0.00666844430896971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,51200,0.4475617938571506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,128,0.0052622221410274506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,8192,0.09502577781677246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,8192,1.6177742216322157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,16384,0.03152622116936578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,6144,0.015271110667122735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,8192,1.0571084552341037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,2560,0.16227289040883383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,2560,0.029008888536029395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,512,0.008970666262838576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,768,0.0069146665434042616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,64,0.008791111409664154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,3072,0.037668443388409086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,768,0.005818666683302985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,51200,0.08681955602433945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,5120,0.011691555380821228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,7168,0.12942666477627224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,3584,0.026309332913822595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,768,0.01889511115021176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,51200,10.603602939181858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,51200,0.018651554981867473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,2048,0.1305333375930786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,512,0.011481777661376528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,1536,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,1536,0.17747910817464194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,32,0.010065777434243096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,16384,0.006037333359320958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,16384,0.04143111242188348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,51200,6.285216013590495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,10240,0.09059555663002862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,4096,0.03869422276814779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,768,0.008076444268226624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,4096,0.11833688947889541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,5120,0.1298675537109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,32,0.00888177752494812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,3584,0.01161688897344801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,3584,0.012108444339699216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,12288,0.06026666694217258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,6144,0.02324177821477254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,7168,4.0025583902994795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,32,0.007559110720952352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,768,0.010228444304731157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,2048,0.030233777231640283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,6144,0.011732444167137146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,64,0.003375111147761345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,128,0.00480000012450748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,3584,0.1357591152191162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,64,0.0068400001360310455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,768,0.006110222389300664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,51200,0.8806657791137695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,8192,0.12180444929334854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,512,0.004304000073009067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,51200,0.16561688317192927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,2048,0.019161777363883126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,65536,0.1397093269560072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,256,0.013032889200581444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,2560,0.011537777880827585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,65536,0.42696089214748806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,2560,0.10306577550040351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,768,0.0064151109092765385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,512,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,512,0.006822222222884496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,512,0.008378666308191087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,3072,0.2348355452219645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,3584,0.13977689213222927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,2560,0.011222222612963783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,1024,0.005665777872006099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,10240,0.0722915530204773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,4096,0.03493600090344747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,5120,0.01257155587275823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,768,0.07631999916500516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,256,0.01570222278436025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,2048,0.010220444036854638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,51200,0.8703902032640246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,64,0.005010666532648934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,128,0.014524444937705994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,512,0.006225777582989798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,65536,0.9768159654405383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,3584,0.03173066510094537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,5120,1.037965350680881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,51200,0.12682667043473986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,4096,0.22569778230455187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,16384,0.4559413592020671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,128,0.004488888714048597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,256,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,12288,2.3205608791775174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,8192,0.11359555191463894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,7168,0.3293208811018202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,256,0.015451555450757345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,1024,0.00702400008837382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,16384,2.1766009860568576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,128,0.04137688875198364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,4096,0.07993777592976888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,8192,0.04982577761014303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,768,0.2581004566616482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,768,0.0036720000207424164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,16384,0.2816195487976074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,3584,0.008008889026112026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,16384,0.05659111340840658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,7168,0.011169777976142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,512,0.005523555394675996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,1536,0.009280000295903947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,51200,0.5520648956298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,128,0.008536888493431939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,32,0.006676444576846228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,2048,0.12240089310540093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,8192,0.03187911046875848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,3584,0.04657511247528923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,65536,0.1536782185236613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,10240,0.028706666496064927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,7168,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,65536,3.1208604176839194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,12288,0.16853600078158906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,12288,0.02810844447877672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,768,0.006559111177921295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,12288,1.22050846947564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,51200,0.2653075589074029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,512,0.007332444190979004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,3072,0.9617626402113172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,2048,0.014202665951516895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,1024,0.015643555257055495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,256,0.003549333247873518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,5120,0.011815110842386881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,51200,0.18902577294243705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,1536,0.00620888877246115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,12288,0.029056890143288508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,2048,0.007726222276687622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,64,0.0027386666172080566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,65536,0.10984710852305095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,64,0.013951111170980664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,3072,0.015632889337009855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,5120,0.0677679975827535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,10240,0.4888604482014974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,16384,0.3259689013163249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,3072,0.01386044422785441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,1024,0.020094222492641874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,256,0.01977511081430647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,2048,0.006846222198671765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,2048,0.010227555202113258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,2560,0.011029333704047732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,2048,0.03351377778583103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,65536,0.07659910784827338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,4096,0.009219555391205682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,1024,0.027794665760464136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,7168,0.0282008879714542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,10240,0.43518755171034074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,32,0.006058666855096817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,32,0.003079111170437601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,8192,0.014179555906189812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,768,0.006058666855096817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,3072,0.043975111511018544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,10240,0.15338667233784994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,1536,0.021134222547213238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,512,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,10240,0.315353790918986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,3584,0.020408888657887776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,10240,0.17166933748457167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,4096,0.047641777329974704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,2560,6.533263312445746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,64,0.0034462221794658234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,12288,0.029414223300086126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,2560,0.18685600492689344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,5120,0.034649779399236046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,6144,0.07725244760513306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,7168,0.12758133146497938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,51200,23.377268473307293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,5120,0.8881768650478787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,1536,0.019350222415394254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,65536,0.14955911371443006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,65536,0.9109840393066406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,10240,0.04827910992834303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,256,0.007400000260935889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,2048,0.011152889165613385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,8192,0.03173777792188857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,128,0.005515555540720622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,1024,0.004467555632193883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,512,0.006585777633719974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,12288,0.3157893286810981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,512,0.025153777665562097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,3072,0.009786666267448002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,2048,0.011429333024554782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,32,0.0052977779673205475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,5120,0.01388177772363027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,51200,1.7572933832804363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,12288,0.0671831104490492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,65536,1.9713412390814886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,128,0.0033724444607893624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,32,0.009512888888518015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,2048,0.14607911639743382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,128,0.004830222162935468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,64,0.00546577779783143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,16384,0.5678995450337728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,512,0.014257777068350049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,3584,0.01012711144155926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,256,0.0042133331298828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,16384,0.22472445170084634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,768,0.031369778845045306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,51200,2.0536702473958335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,5120,1.058744854397244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,64,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,2048,0.07321155733532377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,128,0.016599110431141324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,3072,0.049110223849614464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,3072,0.13824000623491076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,2560,0.40931023491753477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,51200,0.31139998965793186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,65536,0.020991999242040846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,256,0.005099555684460534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,8192,1.1903759638468425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,4096,0.025290666355027094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,3584,0.01312444441848331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,768,0.018792889184421964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,2560,0.03176533182462057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,10240,0.01991911066903008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,10240,0.03492177857293023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,8192,0.09302844603856404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,16384,0.06426578097873263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,3072,0.01366133325629764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,3072,0.07308266560236613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,128,0.12855288717481825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,2048,0.042035556501812406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,2048,0.008809777597586313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,10240,0.277021328608195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,16384,0.39313422309027773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,1536,0.007256888680987888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,7168,0.5944346851772732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,4096,0.011237333218256632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,8192,0.02084355552991231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,7168,0.026329777306980554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,768,0.009140444298585257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,7168,0.0462026662296719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,7168,0.12347111437055801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,2048,0.406529770957099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,51200,0.11078044441011216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,4096,0.0063324446479479475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,128,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,16384,1.921420415242513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,65536,0.14738310707939997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,7168,0.1813911067114936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,7168,0.1168631050321791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,512,0.32074401113722056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,4096,0.0063866666621632045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,512,0.010142222046852112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,51200,0.09420888953738743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,6144,0.010009778042634329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,5120,0.28687556584676105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,64,0.12230932712554932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,12288,4.162825690375434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,12288,1.8486480712890625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,3584,0.029857777886920508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,4096,0.13744444317287868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,8192,0.12498311201731364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,1024,0.021556443638271753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,32,0.002696000039577484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,65536,0.40574844678243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,1536,0.04356088903215197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,768,0.02923111120859782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,32,0.017147555947303772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,65536,5.01914299858941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,51200,0.00925511121749878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,12288,0.03691555394066705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,32,0.007359111474619971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,8192,6.659016927083333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,8192,0.28133511543273926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,2560,0.008071111308203803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,512,0.0645244452688429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,1024,0.004256000121434529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,4096,0.11239377657572429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,2560,0.2642924520704481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,5120,0.2349466747707791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,32,0.0034204444123639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,3072,0.015341333217091031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,1024,0.07397422525617811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,7168,0.026928888426886663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,3072,0.01574844453069899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,8192,0.019009777241282992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,3072,0.010021333065297868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,3584,0.019924443629052904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,1024,0.004056888735956616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,128,0.7547555499606662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,51200,0.05461244450675117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,5120,0.02608799934387207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,4096,0.01418133411142561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,32,0.0041235557032956015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,8192,0.0938826666937934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,8192,0.026280888252788123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,2560,0.04240266813172234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,6144,0.07717155747943454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,2560,0.014114666316244336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,256,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,512,0.003976888954639435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,512,0.21496000554826525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,512,0.006142222219043308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,768,0.03671733207172818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,32,0.03056000007523431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,2560,0.00814222213294771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,65536,0.04432177874777052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,4096,0.012280888855457306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,51200,0.08712533447477554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,2560,0.00739733295308219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,6144,0.02127555509408315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,2048,0.007511110769377814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,7168,0.025133333272404138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,2560,0.04110755523045858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,4096,0.005897777775923411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,256,0.008525333470768398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,7168,0.0660968886481391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,512,0.012843555874294706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,16384,0.07517866955863105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,65536,0.13868178261650935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,65536,0.11164266533321804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,65536,0.8768204583062066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,128,0.0037573331760035623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,16384,0.7789395650227865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,16384,0.0503022207154168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,16384,0.44155022833082413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,64,0.0028702221396896574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,768,0.006335999816656113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,32,0.00979466653532452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,65536,0.11301155885060628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,5120,0.2382897800869412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,4096,0.010862222148312463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,51200,1.6361031002468531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,16384,0.09232978026072185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,1536,0.013544888959990608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,7168,0.01790488925245073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,5120,0.041287110911475286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,5120,0.07199288739098443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,256,0.0058755555914507965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,6144,0.025174222058720056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,1536,0.03470755616823832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,512,0.0032026666320032547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,2048,0.006710222197903528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,512,0.006256888724035687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,3072,0.007654222349325816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,6144,0.04429866539107429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,1536,0.006725333217117522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,16384,0.6470604472690159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,32,0.00905600024594201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,3584,0.698945787217882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,5120,0.012158222496509552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,6144,0.02525955604182349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,7168,0.0485155549314287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,7168,0.09233422411812676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,1024,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,32,0.005159111072619756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,1024,0.023938667443063524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,1536,0.677004443274604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,16384,3.810584174262153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,8192,0.015624889069133334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,7168,0.016948444975747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,2560,0.010294222169452244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,16384,0.0356026656097836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,10240,0.02537155482504103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,6144,0.017064889272054035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,3072,0.2970844374762641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,2048,0.020026667250527274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,8192,0.018090666996108163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,5120,0.009117333425415887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,65536,0.2989271216922336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,65536,0.49131022559271914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,6144,0.01053600013256073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,64,0.0143857780430052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,3584,0.03884622123506334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,2048,0.1339635584089491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,65536,0.18666577339172363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,4096,0.02641333308484819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,10240,0.04247644543647766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,6144,0.006072000082996156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,5120,1.2448551389906142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,12288,0.022823999325434368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,65536,0.19222577412923178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,1536,0.014979556202888489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,256,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,16384,0.08258666594823201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,128,0.003142222141226133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,3072,0.021000888612535264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,2560,0.10006755590438843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,2048,0.005894222193294101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,51200,0.28584355778164333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,16384,0.021670222282409668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,256,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,5120,0.01423111061255137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,32,0.005656888915432825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,4096,0.22620089848836264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,64,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,3072,0.009121778110663096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,7168,0.00656444455186526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,768,0.0513315565056271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,768,0.005899555567238066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,7168,0.18747822443644205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,512,0.005061333378156026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,65536,0.07051377826266818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,256,0.005331555588377847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,65536,0.47341510984632706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,7168,0.024451555477248296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,65536,0.0679688917265998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,6144,0.04429866539107429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,1024,0.012406222522258759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,51200,0.35496266682942706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,16384,0.10053422053654988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,64,0.002700444517864121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,8192,0.013779555757840475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,5120,0.011677333050303988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,128,0.003954666770166821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,512,0.0052168890833854675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,12288,0.017972444494565327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,1536,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,1024,0.003979555434650845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,128,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,7168,0.03324977888001336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,10240,0.38548533121744794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,2048,0.0107004443804423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,2560,0.07858310805426703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,128,0.0026311110705137253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,4096,0.01036533299419615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,10240,0.07280444436603122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,6144,0.013433777623706393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,6144,0.013588444226317935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,6144,0.010052444206343757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,64,0.012061333490742577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,16384,0.030264887544843886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,1536,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,7168,0.042881776889165245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,3584,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,3584,0.8884417745802137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,10240,0.0513973335425059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,5120,0.020407110452651978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,256,0.030639999442630347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,8192,5.448035346137153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,256,0.02679377794265747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,2048,0.007861333588759104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,10240,0.07011110915078057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,51200,3.5618553161621094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,5120,0.16287467214796278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,32,0.009347555538018545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,3584,0.05676977501975166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,51200,0.05883200301064385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,6144,0.006736889067623351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,7168,0.37844000922309023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,51200,1.342010709974501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,6144,0.017702221870422363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,32,0.005404444618357553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,7168,0.048467556635538735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,7168,0.014931556251313953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,2560,0.007947555846638149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,65536,0.5203946431477865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,768,0.003924444317817688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,4096,0.022023111581802368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,768,0.6311235427856445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,16384,0.10053422053654988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,7168,0.01637066735161675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,3072,0.007820444802443186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,1536,0.006816000160243776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,64,0.05197422371970283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,6144,0.10704977644814385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,128,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,1536,0.04000622365209792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,5120,0.9116488562689887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,5120,0.09548711114459568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,3584,0.3592248757680257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,2048,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,16384,0.0392213331328498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,8192,0.17058577802446154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,8192,0.2122195561726888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,32,0.002830222249031067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,768,0.008476444416575963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,2560,0.008018666671382057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,7168,0.01180266671710544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,256,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,512,0.005160000175237656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,7168,0.02216622233390808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,3072,0.03614666726854112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,6144,0.039635555611716375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,12288,0.4547457695007324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,16384,0.1123688883251614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,7168,0.03203999996185303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,256,0.004335999902751711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,64,0.0034506666577524612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,1536,0.010631110933091907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,64,0.0028364445186323593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,3072,0.014608888162506951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,16384,0.047467556264665395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,5120,0.07916800181070964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,1536,0.04287911123699612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,768,0.0069466663731469055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,8192,0.01275111072593265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,32,0.03270577722125583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,1024,0.013754666679435305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,256,0.21978667047288683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,65536,0.01348533315791024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,16384,0.6200444433424208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,768,0.003912888881233004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,51200,0.015559110376569959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,10240,0.015847111741701763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,5120,0.024919999970330134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,128,0.01592177814907498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,5120,0.2696728971269396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,32,0.00435288871328036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,12288,0.015930665863884818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,1536,0.14087466398874918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,6144,0.10104711188210382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,51200,4.422760009765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,12288,0.027152889304690894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,16384,26.665506998697918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,7168,0.02921688887808058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,1536,0.023016000787417095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,4096,0.013694222602579327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,12288,0.017120889491505094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,6144,0.18407111697726777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,4096,0.02257066633966234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,4096,0.1650408903757731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,16384,0.27060622639126247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,16384,0.042246222496032715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,768,0.4088239934709337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,4096,0.010832000109884473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,16384,0.9209271536933051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,128,0.0058222222659322954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,4096,0.010630221830474006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,12288,0.3479288948906793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,2560,0.20396710766686332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,768,0.2662213378482395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,16384,0.04050044549836053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,3584,0.0105262224872907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,8192,0.7841333283318414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,8192,0.06844533152050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,16384,0.01867555578549703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,2048,0.01089333328935835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,768,0.005060444275538127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,32,0.0037546666959921518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,65536,0.06615555286407471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,4096,0.07192089160283406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,5120,0.01683288812637329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,65536,0.06906488868925306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,6144,0.09429511096742417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,6144,0.19734577337900797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,1024,0.00461422238085005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,4096,0.017337777548366122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,32,0.0028986665937635633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,3584,0.019889778561062284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,5120,2.2858568827311196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,2560,0.03313866588804457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,4096,0.014288889037238227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,2048,0.007915555602974361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,256,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,768,0.02235377828280131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,6144,0.020484444167878892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,3584,0.009325332939624786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,32,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,64,0.0038364442686239877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,2560,0.06900622447331746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,1536,0.06274222003089057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,3072,0.2454106542799208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,51200,0.8782586521572537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,16384,0.04232177800602383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,256,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,7168,5.0189289516872835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,16384,0.7144391271803113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,12288,0.05387733379999796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,512,0.008848889006508721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,12288,0.03543466660711501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,1536,0.0406328903304206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,51200,0.015836444165971544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,51200,0.9648737377590604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,16384,0.02725511127048069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,8192,0.020252444677882724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,16384,0.06141510936948988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,1024,0.019638222124841478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,7168,0.0573635564910041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,16384,0.060221334298451744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,4096,0.10057599676979913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,2560,0.029123554627100628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,16384,0.7222115728590223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,256,0.012111110819710625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,512,0.011256888508796692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,2560,0.011172444456153445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,64,0.004302222281694412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,7168,0.2296968830956353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,65536,22.003407796223957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,12288,0.19089867009056938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,5120,0.016501333978441026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,8192,3.95176272922092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,65536,3.0432595147026906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,768,0.0047644442982143825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,5120,0.01591199967596266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,16384,0.5232248836093479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,2048,0.009115555220180089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,2560,0.012886222038004132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,7168,0.011235555840863122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,51200,12.248591105143229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,3072,0.013859555953078799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,7168,0.0869413349363539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,32,0.13006933530171713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,5120,0.3339911037021213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,32,0.12042933040195042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,16384,0.053856889406840004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,16384,0.04080533319049411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,2048,0.10754755470487808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,8192,0.01388088862101237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,2048,0.008828444613350762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,768,0.02510133385658264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,1536,0.0064560001095136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,32,0.1433768934673733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,16384,0.014739554789331226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,1024,0.018831110662884183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,12288,1.54076533847385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,8192,0.008251555263996124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,10240,0.17272888289557564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,6144,0.024633778466118708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,2048,0.03464355402522617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,512,0.04162399967511495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,12288,0.13293866316477457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,64,0.022727110319667395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,128,0.003632889025741153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,5120,0.023277334041065637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,12288,0.42266133096483016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,65536,0.5798311233520508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,8192,0.02201333310869005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,6144,0.03422222203678555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,4096,0.013086222112178802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,3584,0.007870222131411234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,16384,0.2010666661792331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,512,0.010255999863147736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,1536,0.03260711166593764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,8192,0.35991377300686306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,3584,0.007707555260923173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,10240,0.2230586740705702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,10240,0.3010737895965576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,16384,0.012179555164443122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,1024,0.010627555350462595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,1536,0.006899555524190267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,2560,0.012521777715947894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,128,0.004713777866628435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,256,0.007788444558779399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,16384,0.0063999998900625445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,5120,0.18672177526685926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,2560,0.032776888873842024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,4096,0.00814844419558843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,51200,2.5220631493462458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,65536,16.77893235948351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,3072,0.01718577742576599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,7168,0.10856622457504272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,768,0.016227554943826463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,2560,0.010640000303586325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,12288,0.3172791004180908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,4096,0.04424355427424113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,768,0.03183022141456604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,128,0.0265964451763365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,51200,0.1383448839187622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,32,0.020012444920010038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,256,0.00536088893810908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,128,0.009259555902745988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,5120,0.013439111411571503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,51200,0.1347742213143243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,4096,0.06983555687798394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,12288,0.01573955516020457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,64,0.02168088820245531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,32,0.0026337777574857077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,32,0.002671111168132888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,51200,2.2497119903564453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,51200,0.15645599365234375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,128,0.017269333203633625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,32,0.0028844444702068963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,65536,0.6083884239196777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,256,0.004961777892377642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,256,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,5120,0.009844444692134857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,6144,0.30275821685791016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,51200,0.9867386288113065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,768,0.009496889180607265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,5120,0.06044977903366089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,8192,0.02735111117362976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,1536,0.036678221490648054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,512,0.028858668274349634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,768,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,51200,0.10614222288131714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,5120,0.01606666710641649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,768,0.00721244431204266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,3584,0.011963555382357704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,12288,0.04825955629348755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,3584,0.02109333376089732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,128,0.005498666730191972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,12288,0.026407111022207472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,2560,0.1319013304180569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,2560,0.009702222214804756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,51200,0.08142133553822835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,4096,0.2073164383570353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,51200,2.6507350073920355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,8192,0.053336888551712036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,16384,0.03331999977429708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,32,0.007251555720965068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,8192,0.01687733332316081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,7168,0.20867644415961373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,16384,0.23189867867363823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,8192,0.14368711577521429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,16384,0.2643395529852973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,1536,0.00565155554148886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,2048,0.005719999886221356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,3072,0.13340889083014593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,4096,0.013832000394662222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,51200,0.1780266629325019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,16384,0.05267822080188327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,3072,0.01091288857989841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,3072,0.01684799955950843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,2048,0.006806222101052602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,12288,0.22324710422092012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,5120,0.19899822605980766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,16384,0.13920356167687312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,2048,0.014545778433481852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,6144,0.02403555479314592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,1024,0.004695999953481886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,3584,0.1255564424726698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,51200,24.09064229329427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,3584,0.021915555000305176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,3584,0.44565868377685547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,64,0.15266756216684976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,256,0.07868533664279513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,128,0.004607111215591431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,512,0.0033537778589460584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,32,0.0027271111806233725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,4096,0.4501164224412706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,7168,0.062239110469818115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,7168,0.029501332177056208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,1536,0.014102222190962898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,6144,0.20963822470770943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,65536,0.07443910837173462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,1536,0.0068959999415609576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,2048,0.0069155556460221606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,512,0.007683555285135905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,128,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,12288,0.047727111313078135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,768,0.006604444649484422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,65536,0.020925333102544148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,3584,0.045893334680133395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,65536,0.14226310782962376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,64,0.0027431110954946945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,512,0.010643555886215635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,6144,0.011513777905040316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,4096,0.08010222514470418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,65536,40.563574896918404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,16384,0.02453688946035173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,12288,0.01199288914601008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,768,0.10566488901774089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,3584,0.23951821857028535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,5120,0.009453333086437648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,256,0.0030586665703190696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,10240,0.030508445368872747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,32,0.00739733295308219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,7168,0.1764062245686849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,3072,0.009104000197516548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,256,0.002895999906791581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,32,0.04533333248562283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,3072,0.08266844352086385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,5120,0.06652355856365628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,128,0.005320888840489917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,4096,0.013175110850069257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,6144,1.792599148220486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,768,0.0037288889288902283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,512,0.006963555183675554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,2560,0.013209777573744455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,32,0.011759999725553723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,65536,64.70136854383681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,64,0.023064000738991633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,32,0.004831111265553368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,3072,0.1798266702228122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,6144,0.04701688885688782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,12288,0.17321333620283338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,3584,0.01646844380431705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,3072,0.01993955506218804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,768,0.003777777983082665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,256,0.03438755538728502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,65536,0.014690667390823364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,5120,0.010187555518415239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,12288,0.03779911001523336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,64,0.01906488835811615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,7168,0.021956443786621094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,10240,0.026901332868470088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,768,0.007485333416197035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,3584,0.205131557252672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,4096,0.006302222195598815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,1024,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,256,0.00814666681819492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,7168,0.1241253349516127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,6144,0.011186666786670685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,512,0.14494755533006456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,3584,0.08200622267193265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,512,0.03244533472590976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,4096,0.24981511963738334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,512,0.005500444521506627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,65536,0.13691911432478163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,1536,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,51200,0.044574220975240074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,5120,0.01220355596807268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,256,0.023335110810067918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,6144,2.6860035790337453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,12288,0.024353777368863423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,3072,0.046322666936450534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,6144,0.005802666561471091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,51200,0.0828862190246582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,7168,0.02418488926357693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,2048,0.005304000029961268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,12288,0.16749066776699492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,2560,0.06160533428192139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,128,0.007622222105662028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,7168,0.03454666667514377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,1024,0.006022222340106964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,256,0.005366222312053044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,7168,0.3178986708323161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,512,0.008678221868144141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,16384,0.06331466966205172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,16384,0.5885635481940376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,3584,0.02010577751530541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,3584,0.016756445169448853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,7168,3.0368934207492404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,3584,1.2600026660495334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,3072,0.008092444803979661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,16384,4.401288774278429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,51200,0.4032462173038059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,2560,0.008790222307046255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,6144,0.0691502226723565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,128,0.005555555638339784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,65536,0.13573511441548666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,4096,2.871662139892578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,768,0.030415111117892798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,2048,0.041851556963390775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,16384,0.00629066675901413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,4096,0.01683822274208069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,51200,0.16797510782877603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,8192,0.06152977546056112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,512,0.004522666749027041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,128,0.0032142222755485107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,512,0.0047377778424157035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,2048,0.013638222383128272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,10240,0.08306488725874159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,6144,0.007239999870459239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,4096,0.008738666772842407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,2048,0.03606399893760681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,2560,0.014944000376595391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,16384,0.2811768849690755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,768,0.004304000073009067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,2048,0.02104088829623328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,51200,0.05294311046600342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,3072,0.19004621770646837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,6144,0.3874195681677924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,12288,0.11697688367631699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,3584,0.06832800308863322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,768,0.028000000450346205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,128,0.004277333203289243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,5120,0.927662213643392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,65536,0.7378453148735894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,16384,0.8477404382493762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,64,0.16209777196248373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,128,0.014045332868893942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,1024,0.004888888862397936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,5120,0.04494666722085741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,3072,0.009473777479595607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,7168,0.010874666273593903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,1536,0.05806399716271294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,10240,0.1739599969651964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,5120,0.08663466903898452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,5120,0.046037332879172436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,3072,0.20275555716620552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,16384,0.05673066775004069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,2048,0.006903111106819577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,128,0.00519822206762102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,12288,0.07375555568271212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,10240,0.17821510632832846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,5120,0.07279377513461642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,32,0.008222222328186035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,8192,1.2652613321940105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,12288,0.31608178880479604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,2048,0.014681778020328946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,12288,0.1776533391740587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,8192,0.04953866534762912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,8192,0.17286845048268637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,5120,0.48673152923583984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,3072,0.13070578045315212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,2048,0.09211466709772746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,1024,0.004368000146415499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,6144,0.015145777000321282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,1024,0.019182221757041085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,256,0.003629333443111844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,1024,0.11504711045159234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,51200,0.5066720114813911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,512,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,128,0.007582222421964009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,10240,0.06594844659169515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,51200,0.23095554775661895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,1536,0.01479822231663598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,2560,0.07723289065890841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,8192,0.02060622142420875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,256,0.025920889443821375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,2048,0.026160889201694067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,16384,0.02234755622016059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,1536,0.006462222172154321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,7168,0.04764622118737963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,2560,0.011057777537239922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,8192,0.008919999831252629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,12288,0.04071377714474996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,1024,0.017232000827789307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,3584,0.007676444119877285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,12288,0.05905955367618137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,3072,0.1079075535138448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,2560,0.0758497781223721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,2048,0.027464888162083093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,3072,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,3072,0.04142489035924276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,1536,0.02162933349609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,4096,0.0513315565056271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,128,0.0028453332682450614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,51200,0.05033333433998955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,51200,21.410630967881943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,10240,0.09410755501853095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,7168,0.018400889303949144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,256,0.008603555460770925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,51200,0.10084444284439087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,4096,0.05519733495182461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,16384,0.08130844434102376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,128,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,2048,0.039128889640172325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,2560,0.01090666651725769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,512,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,10240,0.02386755579047733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,256,0.029095109966066148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,8192,0.2850080066257053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,1024,0.026446221603287592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,7168,0.021929777330822412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,5120,0.05244444476233589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,8192,0.02570844524436527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,65536,0.09494844410154556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,7168,0.010849778023031024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,5120,0.010632889138327705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,128,0.003055111194650332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,768,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,128,0.007123555574152205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,2048,0.17898310555352104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,12288,0.20471733146243623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,16384,0.024246222443050806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,2560,0.09310666720072429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,256,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,1024,0.009030222064918941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,256,0.0053342220683892565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,128,0.003656889001528422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,6144,0.2611733277638753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,64,0.0037093332244290244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,1536,0.05712088611390856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,12288,0.008014221986134848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,7168,0.12759021917978922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,3584,0.05910399887296888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,5120,0.4080302185482449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,7168,0.021060445242457922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,10240,0.011624889241324531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,10240,0.007811555431948767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,16384,0.1827475494808621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,10240,0.017376888129446242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,768,0.023592889308929443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,1024,0.023024888502226934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,256,0.0028924445311228433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,128,0.014653333359294467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,5120,0.17598933643764922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,2560,0.0121760004096561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,51200,0.31757688522338867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,1536,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,3072,0.015227556228637695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,4096,0.008349333372380998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,4096,0.016747555798954435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,256,0.0033599999215867785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,4096,0.32338044378492564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,1024,0.010037333601050908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,256,0.008942222429646386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,16384,0.028130667077170476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,7168,0.010760000182522668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,6144,0.11016711261537339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,16384,0.3460968865288629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,7168,0.02747111188040839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,512,0.011168888873524137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,64,0.0040755557517210645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,3584,0.006806222101052602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,64,0.002711999954448806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,5120,0.023011555274327595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,12288,0.14597511291503906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,128,0.005872000008821487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,6144,0.08609155813852946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,768,0.00814222213294771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,256,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,4096,0.020729778541458976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,65536,0.07356977462768555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,12288,0.01571466690964169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,3584,0.009568000005351173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,5120,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,7168,0.11372978157467312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,51200,10.34048377143012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,2048,0.008443555898136562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,7168,0.39128888977898496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,3584,0.023041778140597876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,3584,0.010519111322032081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,10240,0.016400888562202454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,8192,0.02455377744303809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,51200,0.041307555304633245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,51200,0.05737688806321886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,1536,0.00721155520942476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,7168,0.42654045422871906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,6144,0.01897333396805657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,32,0.007727999654081132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,32,0.02422044508987003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,8192,0.02519911030928294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,256,0.006063999815119638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,256,0.008788444101810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,512,0.00565155554148886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,512,0.005685333162546158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,32,0.006593777901596493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,64,0.01184266640080346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,768,0.4118817647298177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,1536,0.017795556121402316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,3072,0.09976177745395237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,256,0.008833777573373582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,2048,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,128,0.00546577779783143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,4096,0.008288000192907121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,64,0.11463377210828994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,32,0.004446222136418025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,8192,0.164429333474901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,51200,0.71725066502889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,65536,0.1419395605723063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,256,0.0047271110945277745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,3072,0.028912001185946997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,65536,0.7066791322496203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,1024,0.24306488037109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,7168,0.04784533381462097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,512,0.25663643413119847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,3584,1.4462977515326605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,10240,0.030099554194344416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,1536,0.18306399716271293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,128,0.02165422174665663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,768,0.004842666702138053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,6144,0.6322933302985297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,2048,0.008748444418112436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,8192,0.07465955283906725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,3072,0.00790666706032223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,51200,0.7164959907531738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,64,0.005477333234416113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,16384,0.6339422331915961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,16384,0.028451555305057104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,3072,0.013743999931547375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,1024,0.05886933538648817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,768,0.031014220582114324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,16384,0.05896266963746813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,768,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,32,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,7168,0.6163626776801215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,3584,0.07442222038904826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,6144,0.7019564310709635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,16384,0.011748444702890186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,3072,0.008417777717113495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,16384,0.5066551102532281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,1536,0.01629066632853614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,512,0.04012444284227159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,10240,0.03492710987726847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,256,0.05059911145104302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,2560,0.018057778477668762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,1536,0.028212444649802312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,10240,0.04425866736306084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,10240,0.030172444052166406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,1024,0.013531555732091268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,6144,0.02629866699377696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,16384,0.006439111298984951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,1536,0.007602666815121968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,2560,0.00923911150958803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,1536,0.01425155500570933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,64,0.0076044441925154785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,768,0.03317955467436049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,8192,0.03976000017589993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,1024,0.00509688879052798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,1024,0.0037475555307335327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,7168,0.014191110928853353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,32,0.010556444525718689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,8192,0.011786667009194693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,16384,0.01867555578549703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,3584,0.028487109475665625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,16384,0.03626755542225308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,12288,0.011559999651379056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,768,0.012734221915404001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,10240,0.053892443577448525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,12288,4.149334377712674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,32,0.13203733497195774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,1536,0.006767999794748094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,3072,0.04830044507980347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,16384,0.027449778384632532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,256,0.033504890071021184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,16384,0.015463999576038785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,12288,0.027746667464574177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,1024,0.03683200147416856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,3584,0.07025155756208631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,12288,0.01610311037964291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,32,0.009850666754775578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,64,0.003035555490189128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,2048,0.011747555600272285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,5120,0.016149333781666227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,2560,0.007368889119890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,8192,0.05566222137875027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,4096,0.10256799724366929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,32,0.06005599763658312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,5120,0.09907644324832493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,1536,0.008588444855478074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,256,0.004784000002675586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,3072,0.1536115540398492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,51200,0.05422666668891907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,4096,0.02360711163944668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,8192,0.409518215391371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,1536,0.04757422208786011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,256,0.0036071112586392295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,5120,0.018958222534921434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,2560,0.015948444604873657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,1024,0.011075555450386472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,128,0.00923555592695872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,32,0.005912888795137405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,256,0.004980444494220945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,2048,0.008992888861232335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,8192,0.2968248791164822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,7168,0.1789999935362074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,2560,0.0826079977883233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,256,0.19221956200069854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,512,0.02772444486618042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,6144,0.011892444557613797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,3584,0.012510221865442065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,2560,0.01609511176745097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,10240,1.7565715577867296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,10240,0.016671111186345417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,128,0.0035537779331207275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,51200,6.236858791775173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,6144,0.032200889454947576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,10240,0.10835199885898167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,32,0.01517688896920946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,1024,0.015525332755512662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,10240,0.027662222584088642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,6144,0.3103075557284885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,2048,0.006676444576846228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,6144,0.15816088517506918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,512,0.009885333478450775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,16384,3.409832000732422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,1536,0.00700711127784517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,5120,0.013810666898886362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,64,0.0032924444725116095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,65536,16.98606703016493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,6144,0.044215109613206655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,1024,0.03503111004829407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,512,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,7168,0.13018488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,768,0.03237866693072849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,1536,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,65536,2.0017172495524087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,65536,3.3327484130859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,512,0.010704889065689512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,12288,0.02892000145382351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,8192,0.1595004399617513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,1024,0.013423110875818463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,5120,0.012622222304344177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,8192,0.06891022125879924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,4096,0.44951820373535156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,1536,0.11418755849202473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,3072,0.01423555612564087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,1024,0.019686222076416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,768,0.023237332701683044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,65536,0.11093955569797093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,65536,0.24693422847323945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,6144,0.12876889440748426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,10240,0.06642311149173312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,7168,0.01683288812637329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,2560,0.007902222375075022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,3584,0.008089777496125963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,64,0.032327999671300255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,3584,0.08887555864122178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,7168,0.06321689155366686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,12288,0.010766222245163388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,128,0.0027866665687825945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,2560,0.007159111400445302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,2048,0.01422044469250573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,2048,0.011532444092962476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,10240,0.769809775882297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,64,0.08550755845175849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,768,0.028118222951889038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,6144,0.011242667006121742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,7168,0.27029334174262154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,2048,0.016397333807415433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,3584,0.017768000562985737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,2560,0.018894222047593858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,128,0.0090524446633127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,51200,0.01015822258260515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,128,0.007276444799370236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,2048,0.009308444129096137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,128,0.19388088915083143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,3584,0.09782933526568943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,12288,0.05496977766354879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,128,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,1536,0.3548097875383165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,7168,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,128,0.016508445143699646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,5120,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,65536,6.889661153157552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,2560,0.012503110700183444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,256,0.02195555633968777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,512,0.012168000141779581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,128,0.00553688903649648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,128,0.011552000211344825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,7168,0.010831111007266574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,3072,0.009671111073758867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,3072,0.01296622223324246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,3584,0.020962667134073045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,4096,0.3982035583919949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,2560,0.01907111042075687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,12288,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,2048,0.02365511159102122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,1536,0.16730666160583496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,16384,0.02608266638384925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,1536,0.008434666527642144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,3584,0.09768533044391209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,64,0.003120888852410846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,16384,0.08685688840018378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,8192,0.5070266723632812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,1024,0.025750221477614507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,768,0.04959911108016968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,51200,0.06868800189759997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,512,0.003548444559176763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,1536,0.17929422855377197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,3584,0.03666399916013082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,65536,0.17248445087009004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,8192,0.06462399827109443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,512,0.004925333377387789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,51200,3.4997191958957248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,32,0.02237422267595927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,3072,0.015440000428093804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,512,0.007116444408893585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,8192,0.1512640052371555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,768,0.00978488889005449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,768,0.015802666544914246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,32,0.0052968888647026485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,4096,0.011042666931947073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,1024,0.04273777869012621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,10240,0.4852435323927138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,3584,0.028304000695546467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,3072,0.4809404479132758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,6144,0.17713956038157144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,6144,0.42905865775214297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,3072,0.18513422542148164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,7168,0.012339555554919772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,8192,0.31641244888305664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,5120,0.010773333410422007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,128,0.013560888667901358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,256,0.0030213333666324615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,7168,0.010423111418883005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,51200,0.07051910956700642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,256,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,128,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,768,0.012512000070677863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,768,0.00564177789621883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,2048,0.4635617997911241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,5120,0.01274577776590983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,1536,0.03363200028737386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,16384,0.2407146559821235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,256,0.0029048888633648553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,12288,0.011683555940786997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,4096,0.036610666248533465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,65536,0.06907021999359131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,65536,1.08442136976454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,3584,0.16649066077338323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,16384,0.6628124448988173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,8192,0.03965422179963853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,768,0.007084444165229797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,128,0.005764444255166584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,51200,0.09200711382759942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,2560,0.12018488513098823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,8192,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,2048,0.2133786678314209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,51200,0.013224889006879596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,5120,0.012725333372751871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,5120,0.009871111147933537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,512,0.004447111239035924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,512,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,3584,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,3072,0.007700444095664554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,4096,0.051438222328821816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,512,0.003318222239613533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,2560,0.0092622223827574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,3584,0.018816888332366943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,51200,0.18316355016496447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,3072,0.006216889040337668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,3584,17.456958346896702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,51200,0.21946933534410265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,6144,0.019144000278578866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,65536,0.26712176534864635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,3584,0.013087111214796701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,8192,0.020706666840447318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,65536,0.5706079800923666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,1536,0.05057155423694187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,1536,0.012071111136012606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,64,0.040829333994123675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,1536,0.06959644291136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,1024,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,1536,0.03256266646915012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,3072,0.06686577532026503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,7168,0.017680888374646504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,7168,0.06365688641866048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,12288,0.015055111712879606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,32,0.00518488883972168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,10240,0.18058399359385172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,128,0.005056888692908817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,10240,0.010408889088365767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,2560,0.07867733637491862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,128,0.0028000000036425064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,512,0.0033031110134389666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,32,0.008990222381220924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,1536,0.009995555712117089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,3584,1.8478622436523438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,12288,0.06795022222730848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,1536,0.010773333410422007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,64,0.004707555390066571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,1024,0.01571999986966451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,16384,0.030555556217829388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,4096,0.13278844621446398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,128,0.00489155575633049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,256,0.03266933229234483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,16384,0.23674310578240287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,16384,0.09468266699049209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,768,0.009844444692134857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,10240,0.4298204316033258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,32,0.003034666594531801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,6144,0.022551111049122278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,10240,0.2316462198893229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,4096,0.010253333383136326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,5120,0.021301334102948506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,256,0.010392000277837118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,768,0.009285333255926767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,1536,0.018032888571421306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,8192,0.03386844529045953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,51200,0.5248239835103353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,2560,0.030399110582139757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,1024,0.01200799975130293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,8192,0.058598220348358154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,3584,0.032702220810784235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,4096,0.013358222113715278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,32,0.0053084443012873335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,512,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,768,0.031256887647840716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,6144,0.011354666617181567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,3072,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,12288,0.31441956096225315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,2560,0.009816889133718278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,256,0.002875555513633622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,768,0.006435555716355641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,65536,6.183024088541667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,768,0.007102222078376346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,10240,0.10376266638437907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,12288,0.07372444205813937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,1024,0.0162462227874332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,8192,0.021870222356584337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,2560,0.012727110750145383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,3584,0.06744711266623603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,1536,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,8192,0.07059199942482842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,2560,0.011045333411958484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,2048,0.02217688825395372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,4096,0.011174221833546957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,10240,0.21708533498975965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,768,0.05471200082037184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,4096,0.023057777020666335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,5120,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,10240,0.044212443961037524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,12288,0.015122666954994202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,2048,0.025639999243948195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,768,0.00792711145348019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,2048,0.145069334242079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,3584,9.90120357937283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,128,0.00850755555762185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,2048,0.01443288889196184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,1024,0.00942133367061615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,32,0.004925333377387789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,512,0.01591022147072686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,1536,0.1264666716257731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,512,0.02438666588730282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,8192,0.016436444388495553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,7168,0.024891555309295654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,16384,0.06061066521538628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,64,0.01700711084736718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,7168,0.04167822334501478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,10240,0.024735111329290602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,3584,0.15447289413876003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,6144,0.0516115559471978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,51200,0.5247511333889431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,64,0.004415110995372136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,2048,0.007632888853549957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,16384,1.0993120405409071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,12288,0.1001688904232449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,16384,0.06993066602283053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,51200,35.31289333767361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,65536,0.14256266752878824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,16384,0.0103511114915212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,7168,0.020043555233213637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,128,0.003160888950030009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,10240,0.04681422313054403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,16384,0.03566844595803155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,3584,0.008965333302815756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,3072,0.044405334525638156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,768,0.004837333328194088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,2560,1.228658676147461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,12288,9.227230495876736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,4096,0.09906311167611016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,512,0.011759111450778114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,4096,0.008007999923494127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,4096,0.010790222220950656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,128,0.36081687609354657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,3584,0.3275928762223986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,2048,0.009895111123720804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,2048,0.006793777975771163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,3072,0.00850666645500395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,16384,1.695267571343316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,51200,0.2744595474667019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,65536,3.4015030331081815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,3072,0.060679998662736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,2048,0.02071199980047014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,6144,0.1477528942955865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,64,0.0026791110220882627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,768,0.04378666811519199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,51200,0.712224006652832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,3584,0.01776711146036784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,2560,0.007222221957312689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,8192,0.02349688940578037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,2560,0.008757332960764566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,32,0.002542222125662698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,65536,1.268697738647461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,1536,0.02055288851261139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,7168,0.014841778410805596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,1024,0.04942844311396281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,2560,0.05032444331381056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,512,0.02882666720284356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,1024,0.4190133412679036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,5120,0.010928000013033548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,10240,0.09487644169065688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,5120,0.020480889413091872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,7168,0.01860000027550591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,32,0.13707821898990208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,4096,0.39751821094089085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,5120,0.013960000541475085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,4096,0.08709688981374104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,7168,1.164409743414985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,1536,0.013200889031092325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,7168,0.17325954967074922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,8192,0.15113155047098795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,16384,1.5380178027682836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,64,0.00888533310757743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,256,0.02384888960255517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,4096,0.008433777425024245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,1024,0.029723554849624634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,5120,0.29413866996765137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,5120,0.1233724488152398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,3584,0.007157333195209503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,12288,0.48325157165527344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,64,0.003015999992688497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,51200,0.24595022201538086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,1024,0.011224889092975192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,2560,0.046810666720072426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,7168,0.06446844339370728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,8192,0.011815999945004782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,5120,0.01594666639963786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,64,0.011345778074529437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,32,0.0028008888992998335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,2560,0.01586577792962392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,16384,0.008959111240175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,512,0.0073831114504072405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,7168,0.150200883547465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,3072,0.6003173192342123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,512,0.012698666916953193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,5120,0.013743110828929476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,7168,0.018280888597170513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,12288,0.3919271098242865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,64,0.0038008888562520346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,32,0.00244711111817095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,7168,0.03776799970202976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,256,0.007179555793603261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,8192,0.024950222836600408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,4096,0.16117066807217068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,2560,0.03559111224280463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,32,0.027242667145199243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,8192,0.025480000509156123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,512,0.005577777822812398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,512,0.0075777777367168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,4096,0.019197333190176222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,1536,0.00545155546731419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,10240,0.025219556358125474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,51200,3.9028727213541665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,5120,0.08015377653969659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,8192,0.007461333440409766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,1024,0.007720888488822513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,12288,0.1030835575527615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,6144,0.2565751075744629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,3072,0.048017776674694486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,3584,0.07955555783377753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,2560,0.029532445801628962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,512,0.03400177756945292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,512,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,1536,0.00453333349691497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,6144,0.02544266647762722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,64,0.009698666632175446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,64,0.10372977786593968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,1536,0.010435555544164447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,10240,0.7766026920742459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,2048,0.011371555427710215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,16384,0.3334088855319553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,51200,0.21372444099850127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,65536,0.09741333458158706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,10240,0.48560624652438694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,1536,0.2924373414781359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,128,0.0034053333931499054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,4096,0.07754755682415433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,1024,0.5081457561916775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,51200,0.019955555597941082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,10240,0.009869332942697737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,128,0.0063324446479479475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,10240,0.03422133458985223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,4096,0.010580444501505958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,8192,0.022146667043368023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,1536,0.012139555480745105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,4096,0.01647200021478865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,8192,0.032807999187045626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,128,0.013042666845851474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,128,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,1024,0.045810666349199086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,4096,0.014540443817774454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,64,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,12288,0.23098222414652506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,32,0.002749333365095986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,51200,0.11046222183439468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,7168,0.01629066632853614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,4096,0.21336621708340117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,768,0.09944355487823486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,3072,0.05120977759361267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,8192,0.11123111512925889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,3584,0.012273777690198688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,7168,0.049719111786948315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,10240,0.42174400223626035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,65536,0.2713368998633491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,6144,0.05562311410903931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,128,0.027705777022573683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,51200,1.2046328650580513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,2560,0.07593155569500394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,10240,0.908924420674642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,7168,0.15634666548834905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,12288,0.27833599514431423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,7168,0.018001778258217704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,8192,0.027452444036801655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,3584,0.1585422224468655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,12288,0.025054223007626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,1536,0.03215288784768846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,1024,0.0046640001237392426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,2048,0.010054222411579555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,1536,0.008672888908121321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,3072,0.007052444749408298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,2048,0.08804000086254543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,256,0.029489778810077246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,4096,0.030322667625215318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,8192,0.014398222168286642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,10240,0.022579555710156757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,32,0.0053004444473319584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,8192,0.20123022132449678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,1024,0.05350044369697571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,1536,0.013704000247849358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,128,0.0042444442709287005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,2048,0.014121777481502958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,8192,0.01921333372592926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,16384,0.023019555542204116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,10240,0.2343404557969835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,4096,0.008353778057628209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,256,0.004750222381618288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,51200,0.08831644720501369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,51200,0.2778880066341824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,5120,0.01501066651609209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,10240,0.032050665881898664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,10240,0.5475369029574925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,8192,0.02176711128817664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,768,0.0035217776894569397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,512,0.005624889085690181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,10240,0.08643733130560981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,4096,0.08961688809924656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,128,0.029994666576385498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,51200,0.08552889029184978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,6144,0.031073778867721558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,1536,0.0061928890645504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,256,0.0030257778449190986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,512,0.03018310997221205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,2560,0.04076711005634732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,4096,0.015547555353906421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,32,0.007720000214046902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,32,0.0069253332912921906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,65536,0.07552533017264472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,64,0.0029155556112527847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,1024,0.00406666679514779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,256,0.005174222091833751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,3584,0.016120000018013846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,7168,0.12960267066955566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,1536,0.00980355590581894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,12288,0.01742844449149238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,4096,0.010087110930018956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,3584,0.015769778026474845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,2048,0.0563288860850864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,3584,0.03106400039460924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,10240,0.08006488614612155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,256,0.007328888608349695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,65536,1.055349349975586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,16384,0.5246186786227757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,4096,0.017062221964200337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,2560,0.009512888888518015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,8192,0.11028533511691624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,6144,0.12107200092739528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,65536,3.9908311631944446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,4096,0.7666364245944552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,128,0.006096000058783426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,8192,0.35102399190266925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,64,0.003003555453485913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,5120,0.025747555825445387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,1024,0.020558221472634208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,768,0.020470221837361652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,768,0.01885777711868286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,51200,0.9173289404975044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,128,0.0046826667255825466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,12288,0.02588266630967458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,1536,0.009154666629102495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,3584,0.03902488946914673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,2048,0.010105777945783403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,768,0.01076533314254549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,7168,0.0174666659699546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,256,0.004382222063011593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,16384,0.1382897827360365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,128,0.0027875554644399216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,7168,0.20841777324676514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,7168,0.029123554627100628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,32,0.012842666771676807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,3072,0.07041866911782159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,256,0.0028657778683635923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,8192,0.029145777225494385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,7168,0.09365955326292251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,3072,0.04309244288338555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,1024,0.003979555434650845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,4096,0.06185777982076009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,256,0.011186666786670685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,1024,0.018385777870814007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,64,0.004688888788223267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,2560,0.011740444435013665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,51200,1.9655022091335719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,512,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,3584,0.018960000740157235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,8192,0.047777778572506376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,1536,0.009121778110663096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,7168,0.04307911131117079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,10240,0.36385332213507754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,8192,0.011935110721323224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,7168,0.1382853322558933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,3584,0.0070879997478591064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,16384,0.019164444671736825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,51200,1.055769814385308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,768,0.3455795447031657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,64,0.055897778934902616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,5120,0.018657777044508193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,8192,0.7121751043531629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,768,0.009502222140630087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,32,0.005103999955786599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,7168,0.23522488276163736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,6144,1.8378853268093531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,32,0.008356444537639618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,10240,0.026105778084860906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,5120,0.009454222189055549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,10240,0.13803199927012125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,768,0.010089777410030365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,768,0.013576000101036496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,3584,0.01091822236776352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,1536,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,512,0.0053582220441765254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,768,0.02606933315594991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,768,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,5120,0.2637884351942274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,2560,0.09440000189675225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,4096,0.021207110749350652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,6144,0.13120978408389622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,1024,0.04470400015513102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,1024,0.020419556233617995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,2560,0.01831644442346361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,3584,0.012489777472284106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,768,0.018964444597562153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,10240,0.013623110949993134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,768,0.015318221516079374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,1536,0.006888000087605582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,3072,0.007650666766696506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,768,0.0036959999965296853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,3072,0.18775288263956705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,7168,0.038198222716649376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,128,0.010568000376224518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,1536,1.0699191623263888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,12288,0.675105783674452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,2048,0.5196168687608507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,128,0.006619555668698416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,2560,0.07687199778027005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,51200,0.4208311239878337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,12288,0.325962675942315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,768,0.01421688828203413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,1024,0.003931555483076307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,128,0.004480000171396467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,65536,0.07575377490785387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,32,0.005748444547255834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,16384,0.18943111101786295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,512,0.004456888884305954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,128,0.003863999827040566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,1024,0.03652711047066583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,1536,0.0832408865292867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,7168,0.1688204473919339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,4096,0.03451733456717597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,32,0.005743999862008625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,1536,0.019124444988038804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,128,0.03424533208211263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,3072,0.05920888980229696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,2560,0.010187555518415239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,256,0.08752177821265326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,1024,0.007968000239796108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,10240,0.026025777061780293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,1536,0.00906044410334693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,2048,0.23827732933892143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,256,0.015448888142903646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,768,0.0063902222447925145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,2048,0.017885333961910672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,4096,0.14159822463989258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,12288,0.09484622213575576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,16384,0.6351884206136068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,12288,0.06442488564385308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,32,0.02159999973244137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,768,0.02879022227393256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,6144,0.4872346454196506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,256,0.005069333232111401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,3072,0.009154666629102495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,512,0.0058000000814596815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,10240,0.07165777683258057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,7168,0.024192000428835552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,128,0.020576889316240948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,51200,0.43519287639194065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,16384,0.0339626669883728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,7168,0.36329955524868435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,6144,0.013082666529549493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,3072,0.4336364534166124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,128,0.050795555114746094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,6144,0.018648000227080453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,256,0.00342755557762252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,256,0.0033395555284288195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,256,0.009755555954244401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,51200,0.029776000314288672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,12288,0.015037332971890768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,768,0.02241511146227519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,4096,0.026240888569090102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,7168,0.49111202028062606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,2560,0.017093333933088515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,768,0.0074728892909155945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,3584,0.010392888552612728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,65536,0.07957155836953057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,2560,0.04257066713439094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,5120,0.11198133892483181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,32,0.005106666849719154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,8192,0.00869066682126787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,5120,1.852463192409939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,65536,0.10636711120605469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,5120,0.04325955443912082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,1536,0.024209777514139812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,32,0.004558222161398994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,512,0.02966133422321743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,65536,0.16910221841600206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,16384,10.27726321750217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,1536,0.10510400268766616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,512,0.02516711089346144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,2560,0.05694578091303507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,3072,0.008721777962313758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,5120,0.13510933187272814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,10240,0.14841689003838435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,768,0.005468444277842839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,1536,0.006903111106819577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,10240,0.03727111220359802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,256,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,65536,0.09171644184324478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,512,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,1024,0.04332000017166138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,3584,0.017504000001483493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,32,0.007609777980380588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,65536,0.6407022476196289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,256,0.00572622236278322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,10240,0.014087999860445658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,3584,0.01146044416560067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,7168,0.014440000057220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,32,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,512,0.005777777896987067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,4096,0.009213333328564962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,8192,0.048370665974087186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,7168,0.042131556404961475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,64,0.006821333534187741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,6144,0.01187111106183794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,512,0.06444088617960612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,2560,0.013316444224781461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,1024,0.07693066861894396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,6144,0.06781866815355089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,5120,1.182715521918403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,1536,0.014298665854665967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,768,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,51200,0.11037955681482951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,1024,0.09294844335979885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,16384,0.16846488581763375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,512,0.006143110907740063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,1536,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,65536,0.2962408860524495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,1024,0.02460177739461263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,3072,0.018458666072951425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,1536,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,3584,0.008525333470768398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,16384,0.08693066570493911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,64,0.00258577780591117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,1536,0.22078133953942192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,3584,2.156312942504883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,5120,0.014811555544535318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,3584,0.025786666406525507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,10240,0.04959911108016968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,51200,1.3804373211330836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,16384,0.0292871097723643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,32,0.011327111058764987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,1024,0.01203644441233741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,128,0.171670224931505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,1536,0.007854222423500484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,1024,0.004944889081848992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,8192,0.04210577739609612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,3584,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,16384,0.34873776965671116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,3584,0.11639644040001763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,768,0.008336889247099558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,2560,0.027706666125191584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,2048,0.00849511143234041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,3072,0.05590755409664578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,256,0.17279288503858778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,51200,0.05578844414816963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,2560,0.20768445067935518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,2560,0.007262222468852997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,128,0.17431555853949654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,2560,0.0068711112770769335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,2560,0.14271288447909886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,768,0.021027555068333943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,1536,2.7487057579888234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,1024,0.0052737775776121355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,7168,0.03846755623817444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,4096,0.011908444265524546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,12288,0.0528711113664839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,2048,0.09224710861841838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,128,0.0033084443873829315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,32,0.024622221787770588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,3584,0.02239111065864563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,2048,0.009577777650621202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,2560,3.364088906182183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,6144,0.02602844436963399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,1536,0.0075333333677715724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,3584,0.012299555871221753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,10240,0.05054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,10240,0.0571004417207506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,64,0.11926755640241836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,32,0.003572444534964032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,1024,0.12091822094387478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,4096,0.07964177926381429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,2560,0.01514400045077006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,768,0.005480888817045424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,5120,0.016484444340070088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,10240,0.0162462227874332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,1024,0.006151999864313338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,512,0.21713244915008545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,768,0.007711110843552484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,768,0.003654222107595868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,1536,0.05640266338984171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,2048,0.10033866431978013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,64,0.004546666724814309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,256,0.0031999999450312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,7168,0.010683555569913654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,6144,0.00980888886584176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,12288,0.009497777455382878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,8192,0.19225154982672799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,7168,0.026231999198595684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,6144,0.016294222739007738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,2048,0.005508444375462002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,32,0.0074942219588491656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,32,0.00309244439833694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,3584,0.011155555645624796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,65536,0.08576622274186875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,32,0.04528533418973287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,12288,0.31272978252834743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,256,0.01222222215599484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,1024,0.01258666647805108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,1536,0.04030133287111918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,7168,0.028505778974956934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,12288,0.2543111112382677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,5120,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,5120,0.18667377365960014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,1536,0.004519999855094486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,2560,1.2851875093248155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,12288,0.06960355573230319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,5120,0.2067884471681383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,3584,0.009767110976907942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,7168,0.23300533824496797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,7168,0.014340443743599786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,51200,0.15179288387298584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,4096,0.4668862024943034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,8192,0.04855644371774462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,8192,0.10861866341696845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,64,0.029930667744742498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,512,0.00795022232664956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,16384,0.019874667127927143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,1536,0.00962577760219574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,3072,0.016775111357371014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,1024,0.005535999933878581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,64,0.003387555480003357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,512,0.011419555379284753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,12288,1.3704533047146266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,12288,0.4037928846147325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,7168,0.010416000253624385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,6144,0.02834222382969326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,65536,7.347216288248698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,6144,0.013030221892727746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,16384,0.15965244505140516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,65536,0.5814720259772407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,2048,0.0747493306795756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,3072,0.006822222222884496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,8192,0.027077333794699773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,3072,0.03435466686884562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,10240,0.03596177697181702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,128,0.004606222112973531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,2560,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,65536,0.43151556120978457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,256,0.003444444388151169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,3584,0.02125777800877889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,51200,0.1551991171307034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,32,0.026022222306993272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,12288,0.02385866641998291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,3072,0.016391999191708036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,2560,0.12065777513715957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,5120,0.33869155248006183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,10240,0.024550222688251074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,12288,0.041004445817735456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,16384,0.026535999443795946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,64,0.0028862222615215513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,1536,0.015285332997639975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,512,0.003571555432346132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,128,0.016018667154841952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,32,0.22968355814615884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,6144,0.036087999741236367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,6144,0.02073777715365092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,3072,0.00921688891119427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,64,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,512,0.009768889182143742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,7168,0.41508976618448895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,51200,0.06820622417661878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,65536,0.08548977639940049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,12288,1.4781244066026475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,12288,0.011319110790888468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,12288,0.27804443571302623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,32,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,16384,0.020224889119466145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,6144,0.17746133274502227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,3072,0.02405866649415758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,1536,0.013942221800486246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,1024,0.006021333237489064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,32,0.005802666561471091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,4096,0.080530669954088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,5120,0.07931822538375854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,6144,0.10442044337590535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,128,0.0038951109680864546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,8192,0.39503200848897296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,51200,0.10828177796469794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,1536,0.006439111298984951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,2048,0.006903111106819577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,5120,0.3376933203803168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,1024,0.48003821902804905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,8192,0.21393956078423393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,3072,0.011420444481902652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,512,0.00655377780397733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,2560,0.04791288905673557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,768,0.005816000203291575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,7168,0.01866666641500261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,512,0.009411555197503831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,256,0.0064604443808396654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,8192,0.16469421651628283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,10240,0.049661331706576876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,7168,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,16384,0.10577333635754055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,4096,0.08761066860622829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,256,0.025320000118679468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,2560,0.01850311126973894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,2048,0.02891555428504944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,768,0.08217244678073458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,1536,0.01070844464831882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,65536,0.7901013162400988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,3072,0.007095110913117726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,512,0.003343111111058129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,2560,0.006818666640255187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,1536,0.05312444435225593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,256,0.014831999937693277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,5120,0.620873769124349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,7168,0.007376888559924231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,32,0.007732444339328342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,65536,1.529476483662923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,6144,0.06572888957129584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,16384,0.24744889471266004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,768,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,16384,3.9405411614312063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,128,0.0064737776087390045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,10240,0.006711999989218182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,256,0.15392177634769016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,4096,0.01260444439119763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,4096,0.230200899971856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,3584,0.02934488985273573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,32,0.006861333217885759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,51200,0.03751111030578613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,6144,0.13736089070638022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,6144,0.2029564380645752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,2560,0.00977955593003167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,256,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,256,0.01588888963063558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,32,0.003917333152559069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,3072,0.03363110952907138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,6144,0.013394667042626275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,128,0.0029120000286234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,7168,0.2644435564676921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,1024,0.36002667744954425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,16384,0.4356666670905219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,4096,0.01864355636967553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,5120,0.043268445465299815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,3072,0.08149155643251207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,5120,0.014500444134076437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,8192,0.5765626695421007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,128,0.008385777473449707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,10240,1.9308683607313368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,5120,0.07866488562689887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,51200,3.7876658969455295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,64,0.2588764561547173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,7168,0.0069102222720781965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,5120,0.21962577766842314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,16384,0.025549333956506517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,2560,0.016543111867374845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,256,0.003101333354910215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,512,0.030989332331551447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,256,0.005943111247486538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,128,0.0036391110883818734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,16384,0.013380444712109037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,6144,0.012791111237472959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,12288,0.02306933369901445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,3584,0.11092355516221787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,65536,0.3254835605621338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,8192,0.05520622266663445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,4096,0.10809155305226643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,7168,0.11697422133551703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,64,0.1445582177903917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,64,0.003850666599141227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,5120,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,64,0.014022222823566861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,12288,1.307532416449653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,6144,0.5457253456115723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,10240,0.05784444676505195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,8192,0.0063893331421746155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,10240,0.056483553515540234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,4096,0.034640888373057045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,7168,0.06581155459086101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,10240,0.06714755296707153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,1024,0.0052168890833854675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,10240,0.03757866554790073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,10240,0.2082080046335856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,8192,0.023718222975730896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,32,0.007352888584136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,10240,0.0362026658323076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,3072,0.007520889242490132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,10240,0.11909155050913493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,256,0.047334220674302846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,64,0.005489777773618698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,64,0.0025680000997251934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,5120,0.18919644090864393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,6144,0.020573332905769348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,8192,0.3817982143825955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,128,0.0867697795232137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,768,0.011141333315107556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,128,0.0032791110376516977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,3072,0.08814044131173028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,768,0.04439199964205424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,51200,0.3550613456302219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,128,0.013423999978436364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,7168,0.24277067184448242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,32,0.02640977833006117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,256,0.39999111493428546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,16384,0.07030844688415527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,32,0.009082666701740688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,2048,0.03241511185963949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,2048,0.05309955610169304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,2560,0.04588977826966179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,256,0.007072000039948358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,12288,0.04330133398373922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,256,0.004455111092991299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,7168,0.031855109665128924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,51200,0.11039911376105414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,3072,0.0158595558669832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,3072,0.007814221911960179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,65536,0.2688808970981174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,64,0.005166222237878376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,64,0.13185511695014104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,8192,0.03958666655752394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,51200,0.46109241909450954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,128,0.005763555566469829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,1024,0.2769235504998101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,16384,0.025506666964954797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,12288,0.10029511319266425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,256,0.006531555619504716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,2560,0.011868444581826528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,12288,0.019861333900027804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,6144,0.1702995565202501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,1024,0.03047910994953579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,10240,0.17603822549184164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,2560,0.007382222347789341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,512,0.03260533346070184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,32,0.002599999929467837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,51200,0.0536097784837087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,65536,0.3101671006944445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,2560,0.10912888579898411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,6144,0.21957155068715414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,16384,1.1597608990139432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,3584,0.13782933023240831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,512,0.016696888539526198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,8192,0.010844444235165914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,1536,0.5233893394470215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,6144,0.01830222209294637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,128,0.0032791110376516977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,51200,0.27186044057210285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,2048,0.010123555858929953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,2560,0.030850665436850652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,65536,47.192121717664925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,51200,3.3413325415717234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,256,0.006243555496136348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,32,0.003671110918124517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,8192,1.3136924107869465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,128,0.004361777669853634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,12288,0.18588266107771131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,1536,0.04772888951831394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,1536,0.00961599995692571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,1536,1.1003057691786025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,65536,0.5956133206685384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,3584,0.2771200074089898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,4096,0.00982488857375251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,3072,0.007346666521496243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,51200,0.10195910930633545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,3584,0.026368000441127356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,6144,0.1336488856209649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,512,0.06707200076844957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,2048,0.012656889028019376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,3584,0.0107022225856781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,10240,0.04785511228773329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,8192,0.23712354236178926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,1024,0.006331555545330048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,768,0.022145777940750122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,51200,0.12047378222147624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,7168,0.03450311223665873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,8192,0.035522666242387556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,256,0.007224889265166388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,6144,0.018606222338146634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,7168,0.18897510899437797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,12288,4.992068396674262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,3584,0.015799999237060547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,4096,0.12791200478871664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,4096,0.011404444773991903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,768,0.247016880247328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,16384,0.039777778916888766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,51200,0.6804515520731608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,10240,0.022288888692855835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,51200,1.7729191250271266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,12288,0.09713600079218547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,256,0.004026666697528628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,2048,0.04477777745988634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,3584,0.06343289216359456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,3072,0.011267555256684622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,3072,0.009831111464235518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,6144,0.07105244530571832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,256,0.00602311102880372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,2560,0.08010222514470418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,2560,0.02612888813018799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,4096,0.10840355687671238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,3584,0.038542220989863075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,51200,0.08580444256464641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,2560,0.01274311128589842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,256,0.010803555448849997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,6144,0.016924444172117446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,12288,0.03218044506178962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,256,0.0035137778355015647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,3584,0.007982222570313348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,10240,0.20031111770206025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,51200,0.6941057840983073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,1536,0.01016711112525728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,1536,0.025871111287011042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,512,0.023974221613672044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,1536,0.008496888809733922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,1536,0.05174755718972948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,7168,0.2892844412061903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,768,0.005653333332803514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,1536,0.007465778125656976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,12288,2.7737945980495877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,65536,0.10713155402077569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,2560,0.01531733406914605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,4096,1.7896675533718531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,10240,0.05042044321695963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,7168,0.02936089038848877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,12288,0.058146668805016406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,8192,0.6809200180901421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,10240,0.01681600014368693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,12288,0.3934008810255263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,1024,0.04589688777923584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,51200,2.987899568345812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,3584,0.016072000066439312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,1024,0.016241777274343703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,51200,45.430131700303825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,8192,0.05455466773774889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,10240,0.010484444598356882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,4096,0.0829111138979594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,8192,0.05066133207745022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,64,0.033033778270085655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,3584,0.010957332948843638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,3072,0.8477173381381564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,7168,0.014607111612955729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,256,0.006829333388143116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,3584,0.026126222478018865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,12288,0.21358933713701037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,32,0.013712888790501488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,8192,0.010566222170988718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,12288,0.007978666987684038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,51200,0.058131555716196694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,1024,0.046078221665488354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,5120,0.15073600080278185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,16384,0.05657688776652018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,65536,6.121592203776042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,2048,0.017871111631393433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,32,0.0034462221794658234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,32,0.10632533497280544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,5120,0.018434666925006442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,768,0.012801777985360889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,8192,0.15248178111182317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,3072,0.010755555497275459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,1024,0.004088888979620404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,5120,0.13548444377051458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,3584,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,512,0.021617778473430212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,12288,0.01610933409796821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,128,0.0042568888101312844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,512,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,5120,0.021945777866575453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,128,0.004222222086456087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,6144,0.006050666587220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,128,0.00499911109606425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,1536,0.021912889348136053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,7168,0.013055110971132914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,5120,0.01148888882663515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,6144,0.15642666816711426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,16384,0.23871111869812012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,65536,7.323550754123264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,512,0.03382488754060533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,2048,0.06740088595284356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,16384,0.04155377878083123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,64,0.005452444569932089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,2560,0.017285333739386666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,2048,0.0057555557125144536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,2048,6.2804383171929254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,65536,0.18938844733768037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,12288,1.3068346447414823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,4096,0.007409777906205919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,6144,0.01458044515715705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,128,0.009053332938088311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,7168,0.08511555857128567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,512,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,10240,0.12200978067186143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,256,0.004648888690604104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,2048,0.010275555153687796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,51200,0.3673146565755208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,5120,0.025993777645958796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,10240,0.04732444551255968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,5120,0.024217777782016333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,4096,0.01108266661564509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,128,0.0882177750269572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,6144,1.5369164148966472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,3584,0.2876639895968967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,1536,0.00870311094654931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,768,0.004561777744028303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,6144,0.0319164428446028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,4096,0.0310204459561242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,3072,0.07296622461742826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,3584,0.07516444391674466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,2560,0.029267556137508813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,32,0.05830133623547024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,8192,0.6396631134880914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,7168,0.34099200036790633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,32,0.012759110993809171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,7168,0.012997333374288348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,3584,0.028537776735093858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,64,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,6144,0.012540444731712341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,51200,0.05784444676505195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,2560,0.06625066863165961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,3072,0.006583111153708563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,2048,0.008009778128729926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,3072,0.008819555242856344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,768,0.015196444259749519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,64,0.718281798892551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,51200,0.27304622862074107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,3072,0.08278044727113512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,6144,0.01720711092154185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,32,0.002985777747299936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,3072,0.04563200142648485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,2048,0.06502844227684869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,8192,0.008822222550710043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,128,0.011434666812419891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,8192,0.07121155659357707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,256,0.0052693333062860705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,4096,0.030067556434207495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,2048,0.005485333502292633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,6144,0.10903111431333755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,32,0.005517333332035277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,256,0.021459556288189355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,3072,0.006635555376609166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,5120,0.03449777762095133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,12288,0.0315688881609175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,64,0.005065777649482091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,256,0.021299555897712708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,2048,0.009232000344329411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,1024,0.01075466639465756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,32,0.00655644428398874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,1024,0.04068622324201796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,32,0.002605333303411802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,3584,0.011775111158688864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,12288,0.030724445978800457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,5120,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,8192,0.26279645495944554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,5120,0.34084534645080566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,128,0.005838222387764189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,10240,0.03696711195839776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,2560,0.006635555376609166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,7168,0.01110577748881446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,256,0.004604444321658877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,32,0.019455111689037748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,4096,0.04794844322734409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,6144,0.8742808765835233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,8192,0.21178666750590006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,64,0.00254488881263468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,65536,3.8432214525010853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,128,0.014515555567211576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,3584,0.02863022353914049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,65536,0.8468746609157987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,128,0.0052782222628593445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,768,0.04606399933497111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,3072,0.04841599861780802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,64,0.007713778151406183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,51200,0.06846933232413398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,12288,0.017920000685585868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,3584,0.06651110781563653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,512,0.005937777873542573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,12288,0.14056089189317492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,128,0.007733333441946242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,10240,0.006899555524190267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,1536,0.016223111086421542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,512,0.0034515555534097883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,12288,0.020872889293564692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,64,0.008268444074524773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,7168,0.2308008935716417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,2048,0.007910221815109253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,51200,0.33957333034939235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,10240,0.8483289082845052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,8192,0.023221333821614582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,512,0.04315733247333103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,512,0.003721777763631609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,768,0.010784000158309937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,6144,0.01646933290693495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,4096,0.3017511102888319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,6144,0.013583111266295115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,8192,0.7196533415052625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,64,0.003640888879696528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,1024,0.007647111184067196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,4096,0.06282489167319404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,65536,0.01680799987581041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,51200,0.23342490196228027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,6144,0.042135109504063926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,8192,0.11276533868577744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,5120,0.012872888810104795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,6144,0.025524444050259058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,3072,0.05560355716281467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,3584,0.006807111203670502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,768,0.0064106666379504735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,3584,0.058845334582858615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,3072,0.010326222413116032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,2048,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,256,0.004117333226733738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,1536,0.027842667367723253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,12288,0.016234666109085083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,32,0.0839315586619907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,256,0.009903999666372934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,1024,0.012331555287043253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,65536,0.04501333170466953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,32,0.0047475554876857335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,768,0.011018666956159802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,7168,0.18643555376264784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,2048,0.2642853260040283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,12288,0.037568890386157565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,1024,0.02088266611099243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,65536,0.07407910956276788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,1024,0.009072889056470659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,10240,10.05764431423611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,65536,0.20251199934217665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,3584,0.010856000085671743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,32,0.004864888886610667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,3072,0.007807999849319458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,1024,0.01588000026014116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,3584,0.1321644385655721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,128,0.19571022192637125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,8192,0.05628266599443224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,51200,2.8322870466444225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,10240,0.14559111330244276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,128,0.026855111122131348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,2048,0.0185199992524253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,3072,0.026477333572175767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,7168,0.010234666367371878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,2048,0.16046755843692356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,16384,0.35528087615966797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,8192,0.06498044729232788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,768,0.02159022291501363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,6144,0.02609688871436649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,2560,0.006019555446174409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,128,0.007048889166778988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,32,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,1024,0.02594311038653056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,4096,0.061475555102030434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,768,0.015304888288180033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,65536,0.1217599974738227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,2048,0.0166640000210868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,768,0.029976000388463337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,12288,0.010260444548394945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,4096,0.02257333364751604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,128,0.005019555489222209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,5120,0.11667111184861924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,3072,0.13946578237745497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,512,0.0034640000926123727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,5120,0.1055377788013882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,65536,0.08358222246170044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,512,0.005607999861240387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,256,0.005604444278611078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,3072,0.059438221984439425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,4096,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,256,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,256,0.011529777612951068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,12288,0.06670133272806804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,128,0.0058471109304163195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,768,0.014494222071435718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,256,0.01904000010755327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,51200,0.1635164419809977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,32,0.007635555333561367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,128,0.0026435556097163092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,12288,0.023377777801619634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,2048,0.010556444525718689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,12288,0.04274222254753113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,2048,0.008715555899673039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,3584,1.2925910949707031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,768,0.024414221445719402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,512,0.004459555364317364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,3584,0.018457776970333524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,16384,0.03419555558098687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,8192,0.15023733509911433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,12288,0.0351493325498369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,10240,0.006631111105283101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,4096,0.9253688388400607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,5120,0.016002666619088914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,3072,0.025289777252409194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,3072,0.04629955689112345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,512,0.04054044352637397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,10240,1.2458648681640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,256,0.053520000643200345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,6144,0.08630577723185222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,51200,2.257764392428928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,256,0.006365333166387346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,6144,0.06688177585601807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,12288,0.011054221954610614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,65536,3.0479191674126516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,7168,0.01713511182202233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,256,0.0032515554792351196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,256,0.01828177769978841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,1536,0.00535999983549118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,51200,0.08352088928222656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,8192,0.14331555366516113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,51200,0.09652711285485162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,4096,0.013346666263209449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,1024,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,51200,0.3380897839864095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,10240,0.008553778131802877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,10240,0.03519377774662442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,128,0.008340444829728868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,256,0.017802667286660936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,768,0.037704888317320086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,10240,0.4176231225331624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,1536,0.009116444322797988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,16384,0.0921635561519199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,5120,0.1338284413019816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,3072,0.12179288599226211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,768,0.021720000439220007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,12288,1.2777786254882812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,2560,0.011398221883508893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,256,0.0121760004096561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,64,0.02408355474472046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,2048,0.038049777348836265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,3072,0.008681777450773451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,8192,0.803466690911187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,8192,0.30690132247077095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,5120,0.011122666299343109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,64,0.1106426649623447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,8192,0.47795465257432723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,256,0.019498666127522785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,512,0.003310222178697586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,128,0.004741333425045013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,2560,0.013561777770519257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,6144,0.08277066548665364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,128,0.006311111152172089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,1536,0.034019556310441755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,7168,0.0118222220076455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,7168,0.06742578082614474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,8192,0.10204444328943889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,1024,0.021527111530303955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,2048,0.006751110984219446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,32,0.010751999914646149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,16384,0.7406889067755805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,1024,0.01832444469134013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,65536,1.08333863152398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,32,0.008343111309740279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,2560,0.09652622540791829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,3072,0.05652533637152778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,7168,0.11168532901340061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,16384,0.039615111218558416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,65536,0.12097599771287705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,5120,0.009187555147541894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,12288,0.14687821600172254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,16384,0.023708444502618577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,8192,0.2658791012234158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,1024,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,1536,0.05917155742645264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,3072,0.7614569134182401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,64,0.01763199932045407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,7168,0.0074542222751511475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,3584,0.01072622256146537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,10240,0.5507555537753636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,768,0.010760000182522668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,65536,1.8087245093451605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,4096,0.17711199654473198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,32,0.008632889224423302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,12288,0.05285155442025927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,3072,0.1118124458524916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,8192,0.031112889448801678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,2560,0.8049670855204264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,7168,0.012650666965378655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,7168,0.18126222822401258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,3072,0.008386666576067606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,768,0.0037493333220481873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,6144,0.050175110499064125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,4096,0.28968267970614964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,12288,0.19444444444444445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,1536,0.07242400116390653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,6144,0.025145777397685583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,7168,0.0328800015979343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,3072,0.007495111061467065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,2560,0.18605332904391816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,8192,0.31087199846903485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,128,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,4096,0.014857777290874057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,1536,0.011357333097192975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,16384,0.04340177774429321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,3584,0.015273777974976433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,51200,0.26770133442348903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,7168,0.9907866583930122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,16384,4.894122653537326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,5120,0.02633511192268795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,2048,0.007837333612971837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,1536,0.04198399848408169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,512,0.02312711046801673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,512,0.01107022249036365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,5120,0.04085510969161987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,128,0.004279110994603899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,7168,0.04167911079194811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,32,0.012736000120639801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,5120,0.010968888799349466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,2560,0.007289778027269576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,1024,0.020299555526839364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,10240,0.022083555658658344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,1024,0.04896533489227295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,10240,0.018765333626005385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,7168,0.06395644611782499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,10240,0.09158400032255386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,128,0.006141333530346553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,10240,0.01110666659143236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,51200,0.989158206515842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,2048,0.0058462222417195635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,5120,0.0140684445699056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,10240,0.036566221051745944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,4096,0.08329244454701741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,768,0.012433778080675336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,2048,0.006712889091836081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,64,0.004193777839342753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,7168,0.3766782283782959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,12288,0.028946667909622192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,128,0.004560889055331548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,7168,0.033357332150141396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,16384,0.020554666717847187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,51200,0.09531111187405056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,768,0.006210666563775804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,32,0.35089866320292157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,5120,0.18979022237989637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,256,0.008106666306654612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,8192,0.10684444506963094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,7168,0.04292177822854784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,3072,0.05626755290561252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,16384,0.03850488861401876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,10240,0.009927111367384592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,51200,3.815273708767361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,3584,0.1702728933758206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,256,0.0052133335007561584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,1024,0.017849778135617573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,3584,0.36526399188571507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,6144,5.22491455078125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,32,0.0026951111439201566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,3584,0.40299555990431046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,7168,0.015276443627145557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,1024,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,4096,0.09504089090559219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,5120,0.04172177778349983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,8192,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,768,0.04678844412167867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,1536,0.006486222147941589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,3584,0.010009778042634329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,5120,0.009517333573765224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,3584,0.047169778082105845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,7168,0.01349955548842748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,51200,2.8571590847439237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,7168,0.052519109514024526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,2048,0.008166222108734978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,1024,0.05262400044335259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,4096,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,3072,0.012952889005343119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,128,0.005436444448100195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,256,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,128,0.0447324448161655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,512,0.0598542226685418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,768,0.012278222375445895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,3584,0.012883555557992725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,512,0.017363554901546903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,1024,0.3358115620083279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,16384,0.011567999919255575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,64,0.013640888863139682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,6144,0.015204444527626038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,3584,0.07119555605782403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,512,0.005843555761708154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,128,0.022984888818528917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,2048,0.09541866514417861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,512,0.03730666637420654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,2560,0.006204444501135085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,8192,0.017196445001496207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,512,0.06313244501749675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,768,0.004087111188305749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,512,0.01646311084429423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,10240,0.014874666929244995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,7168,0.02767555581198798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,65536,7.959086100260417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,8192,0.4703822135925293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,65536,0.2358622286054823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,1024,0.29245156712002224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,10240,0.018980445133315194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,1536,0.008839111361238692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,512,0.005856888989607493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,3584,0.3256933424207899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,256,0.003141333245568805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,128,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,12288,0.03085511260562473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,51200,1.0560017691718209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,51200,0.05667022201750013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,256,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,4096,0.11515822675493027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,2560,0.011203555597199334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,256,0.0032088889016045462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,768,0.0321048895517985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,8192,0.022100443641344707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,768,0.0058755555914507965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,51200,1.5604995091756184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,16384,0.026428444517983332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,3584,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,10240,0.17919554975297716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,6144,0.026744888888465032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,6144,0.03644444545110067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,16384,0.08939377466837566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,1024,0.0052044445441828836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,4096,0.27553510665893555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,2048,0.013493333425786761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,2560,0.011071999867757162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,2560,0.007696000238259633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,4096,0.0347031123108334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,128,0.004415110995372136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,5120,0.009009777671760982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,8192,0.021548445026079815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,10240,0.03333510955174764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,2048,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,768,0.006468444648716185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,5120,0.5667875607808431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,51200,0.1721822288301256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,10240,0.017802667286660936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,32,0.01477777792347802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,768,0.00702044450574451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,256,0.01553422212600708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,16384,0.0076453329788313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,2048,0.09488533602820502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,6144,0.08475733465618557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,16384,0.018560889694425795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,5120,0.03443110982577006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,256,0.0033200000309281876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,12288,0.019900444481107924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,6144,0.02567377852069007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,64,0.0026293332791990707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,1536,0.015562666787041558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,32,0.005507555686765247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,1536,0.03330933385425144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,10240,0.09122933281792535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,2560,0.10313955280515884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,2048,0.00490133340160052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,51200,0.05785955323113335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,256,0.013459555804729462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,5120,0.043952888912624784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,256,0.0031271111220121384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,2048,0.008760000268618265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,5120,0.010703999963071613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,1024,0.03495644529660543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,64,0.005435555759403441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,7168,0.015263110399246216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,1024,0.005971555494599872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,64,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,4096,0.038513776328828596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,2048,0.512971560160319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,6144,0.015558222929636637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,4096,0.06913244724273682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,8192,0.013247999880048962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,3584,0.024294222394625347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,3072,0.021708443760871887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,1536,0.10533955362108018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,128,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,51200,0.08912711011038886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,3072,0.011136888629860349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,128,0.01924622224436866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,8192,0.013768889009952545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,12288,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,6144,0.02051733268631829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,32,0.0028560000161329904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,6144,0.005945777727497949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,512,0.0038311113086011675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,1024,0.05788177914089627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,2048,0.044328000810411244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,256,0.022479999396536086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,3584,0.010933332973056369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,7168,0.32388443417019314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,2560,0.007104000283612146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,7168,0.020967110991477966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,1024,0.004034666551484002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,16384,0.019684443871180218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,7168,0.013958222336239286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,12288,0.2976124551561144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,8192,0.026449778013759192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,7168,0.02664533257484436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,65536,0.3951999876234267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,65536,0.05750311083263821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,256,0.005704000178310607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,5120,0.03527999917666117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,65536,0.07369244760937162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,8192,0.013392888837390475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,768,0.003524444583389494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,128,0.003301333429084884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,5120,0.04301066531075371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,4096,0.03234222200181749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,512,0.02200088898340861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,1536,0.0062977779242727495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,1536,0.00461422238085005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,768,0.21846755345662436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,256,0.0029724445194005966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,64,0.006991110742092133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,6144,0.03501155641343858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,51200,0.16446933481428358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,6144,0.06466577450434367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,10240,0.028483556376563177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,512,0.3164275487263997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,7168,0.008456889126035903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,10240,0.0417502224445343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,12288,0.04093688726425171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,6144,0.15950844022962782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,7168,0.010635555618339114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,10240,0.20572266313764784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,128,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,10240,0.9812018076578776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,768,0.006628444625271692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,3584,0.18842044141557482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,3072,0.018182223041852314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,51200,1.0925413767496746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,5120,0.19418933656480578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,51200,0.08479555447896321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,512,0.020761777957280476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,128,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,51200,0.05933066871431139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,5120,0.06804355647828844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,7168,0.24783110618591309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,3584,0.01681866745154063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,3584,0.02409422232045068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,6144,1.2286799748738606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,7168,0.009455111291673448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,2560,0.017667555146747164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,768,0.010968888799349466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,51200,0.7505768669976128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,2048,0.041365332073635526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,51200,1.6472675535413954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,8192,0.016677333248986136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,4096,0.023751111494170293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,1024,0.10884000195397271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,4096,1.3644239637586806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,4096,0.013330666555298699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,3072,0.00796888851457172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,256,0.005939555664857228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,128,0.007797333101431529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,512,0.005428444594144821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,5120,0.12605155838860407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,5120,0.007244444555706448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,65536,1.3134924570719402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,64,0.0030097777230872046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,51200,1.6434053844875758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,2048,0.0326275560590956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,2048,0.03888355692227682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,1024,0.004785777793990241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,3584,0.14722932709587946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,12288,0.3208853403727214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,51200,0.9599449369642469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,3584,0.01773866679933336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,2048,0.006604444649484422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,7168,0.04994666576385498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,65536,0.4411991172366672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,2560,0.1100684470600552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,256,0.005600888695981767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,32,0.002887111157178879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,128,0.006759111252095964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,12288,0.0396453340848287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,256,0.005626666463083691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,3072,0.01130844404300054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,8192,0.1277893384297689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,1536,0.047091556919945605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,16384,0.015240889456537036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,51200,0.1256728834576077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,2048,0.010306666294733683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,768,0.012851555314328937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,256,0.003067555526892344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,768,0.010060444474220276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,4096,0.012616888516479068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,3072,0.03841955463091532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,256,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,16384,0.03196000059445699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,1536,0.006440889090299606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,128,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,12288,0.01586844523747762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,51200,0.0754293335808648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,128,0.0033520000676314035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,1024,0.006025777922736273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,8192,0.023034666975339253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,1024,0.007852444218264686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,8192,0.015812445018026564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,51200,0.12074755297766791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,768,0.006414222220579783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,3584,0.022745778163274128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,12288,0.014210666219393412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,3584,0.07041155629687838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,64,0.03110044532352024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,3072,0.010336000058386061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,6144,0.010182222558392419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,3072,0.013071999781661563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,65536,0.3067733446756999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,2048,0.013426666458447775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,7168,0.030849777989917334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,2048,0.0075146663520071245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,1024,0.016023111012246873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,7168,0.0450817777050866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,1536,0.04385777645640903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,51200,0.2963617907630073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,10240,0.03637955586115519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,5120,0.014765333798196582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,12288,1.1699270672268338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,51200,0.2611493269602458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,65536,0.41952000723944766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,1024,0.010709332923094431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,2048,0.007496889266702864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,2048,0.043246222866906055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,2560,0.010127999716334874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,32,0.009280000295903947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,8192,0.24137955241733125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,512,0.009313777916961247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,1536,0.014668444792429606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,65536,0.0996755560239156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,4096,0.11811644501156277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,128,0.0027911110470692315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,8192,0.02198666665289137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,65536,22.585767957899307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,256,0.006873777757088344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,16384,0.045795556571748525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,6144,0.017753778232468497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,512,0.014404444230927361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,2560,0.008170666793982187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,65536,0.6177120208740234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,51200,0.9115582572089301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,32,0.01203733351495531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,3072,0.23611021041870117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,2560,0.08243822389178805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,512,0.01573422220018175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,128,0.05492800143029955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,6144,0.027793778313530818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,8192,0.1039244466357761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,3072,0.011078221930397881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,65536,0.7875822385152181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,768,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,7168,0.06026577949523926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,10240,0.02176355487770504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,7168,0.23797421985202363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,1024,0.005984888722499211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,128,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,32,0.005592888842026393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,6144,0.0140328887436125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,1024,0.005327111317051782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,4096,0.00591466658645206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,768,0.006112888869312074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,16384,0.561988459693061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,2048,0.008852444589138031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,64,0.008310221963458592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,1024,0.017543110582563613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,7168,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,7168,0.6576284302605523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,1024,0.010412444670995077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,10240,0.07149066527684529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,32,0.007869333028793335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,3584,0.010320888625250923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,12288,0.0574586656358507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,1536,0.009679999616410997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,1024,0.005022222383154763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,6144,0.03888266616397434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,4096,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,8192,0.011774222056070963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,64,0.01090755561987559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,128,0.004517333375083076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,4096,0.006490666833188798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,256,0.02160177793767717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,12288,0.10005688667297363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,3584,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,768,0.013434666726324292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,128,0.006613333192136552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,64,0.010317333042621613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,8192,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,2560,0.05154755711555481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,6144,0.21106310685475668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,1024,0.025073778298166063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,10240,0.11339555846320258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,5120,0.24813599056667754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,51200,0.1523911158243815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,512,0.005883555445406172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,12288,0.027288888891537983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,5120,0.11536267068650986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,3584,0.034592890077167086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,6144,0.358695109685262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,3072,0.015600888265503777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,16384,0.0576160020298428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,128,0.04342666599485609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,6144,0.1140400038825141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,5120,0.03541244400872125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,1536,0.048735111951828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,65536,5.721698760986328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,51200,1.1953759723239474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,6144,0.01611022154490153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,1024,0.07434222433302137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,3584,0.012671999633312225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,65536,0.08723911311891343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,512,0.03891288903024461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,1024,0.06934933529959784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,7168,0.01625866691271464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,5120,0.008080888953473832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,512,0.00849777791235182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,3584,0.08518133560816447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,512,0.0036311112344264984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,7168,0.0794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,8192,0.015429332852363586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,5120,3.6151546902126737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,7168,0.03885777791341146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,768,0.011365333365069495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,7168,0.0658444431093004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,12288,0.2685511112213135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,7168,0.16944977972242567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,1536,0.012143111063374413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,1536,0.040017777019076876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,10240,0.07485333416197035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,5120,1.7813733418782551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,512,0.006351111249791251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,3584,0.012390221986505719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,1024,0.0369759996732076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,7168,0.013393777940008374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,5120,0.04538311229811775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,65536,0.14429244730207655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,1024,0.003961777935425441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,12288,0.03559022148450216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,6144,0.01626666718059116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,128,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,8192,0.01718488832314809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,12288,0.06602311134338379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,1024,0.11866311232248943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,8192,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,2560,0.007667555577225155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,32,0.0026577777332729767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,256,0.025795555777019922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,512,0.005646222167544895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,64,0.014719110396173267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,12288,0.016160888804329764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,256,0.003617777592606015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,2560,0.08529688914616902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,256,0.008208889100286696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,8192,0.3159911102718777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,32,0.033350222640567355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,7168,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,12288,0.016230222251680162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,3584,0.08899111217922634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,512,0.012672888735930124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,512,0.005168888717889786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,3072,0.048319111267725624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,256,0.004772444566090901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,128,0.021799110703998145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,10240,0.17545866966247559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,7168,0.013134222063753339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,4096,0.008081778056091731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,768,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,6144,0.010029333333174387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,4096,0.010452444354693094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,5120,0.011947555674446953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,32,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,3072,0.0425182216697269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,10240,0.079821334944831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,256,0.005039111193683413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,256,0.012661333713266583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,768,0.01219733307758967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,2048,0.07104266352123685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,7168,0.015098666151364645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,1024,0.0058151111006736755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,10240,0.3889119890001085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,7168,0.019528888993793063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,256,0.015094222293959724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,5120,0.22277243932088217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,4096,1.354715559217665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,3584,0.019170666734377544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,2048,0.03093866507212321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,768,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,3584,0.06644622484842937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,7168,0.08827733331256443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,16384,0.01445777714252472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,2560,0.03821422325240241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,51200,0.05567377805709839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,12288,0.9601804945203992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,128,0.004017777740955353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,2560,0.03513244456715054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,10240,0.01498044364982181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,12288,0.015107555521859063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,2048,0.007184888753626082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,2048,0.007307555940416124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,5120,0.21750844849480522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,32,0.006613333192136552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,128,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,5120,0.10530489020877415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,128,0.002995555599530538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,2560,0.08561866813235813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,10240,4.547888861762153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,128,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,3072,0.031650665733549334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,65536,0.1987768941455417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,8192,11.300379435221354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,6144,0.016116445263226826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,3072,0.0222044438123703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,1024,0.029954665237002905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,10240,0.026941332552168105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,64,0.010101333260536194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,1024,0.009393778112199571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,3584,0.04124622212515937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,1024,0.004577777865860197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,2560,0.147088885307312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,6144,0.19521955649058023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,16384,0.04122755593723721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,512,0.005666666560702854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,8192,3.106405258178711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,8192,0.10498844252692328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,4096,0.01606044504377577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,256,0.04187022315131294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,768,0.19281689325968424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,1536,0.028032888968785603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,2560,0.05324888891643948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,51200,13.297126770019531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,6144,4.299471961127387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,256,0.010599111517270407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,64,0.02639377779430813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,32,0.007823111282454597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,8192,0.042753779225879245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,65536,0.07620000176959567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,6144,0.0958791110250685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,6144,0.010416000253624385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,3584,0.018339556124475267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,6144,0.019476445184813607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,6144,0.10860266950395371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,64,0.015649777319696214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,768,0.010657777388890585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,12288,0.03787022166781955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,1536,0.06266844272613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,128,0.003034666594531801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,6144,0.01664622293578254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,1024,0.023776888847351074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,6144,0.006799110935793982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,3072,0.27707910537719727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,512,0.007183999651008182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,5120,0.2820604377322727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,4096,0.08301599820454915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,3072,5.936443752712673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,128,0.0028799999919202593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,3072,0.00775022225247489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,10240,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,768,0.03117155697610643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,6144,0.18840532832675508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,256,0.008360000120268928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,6144,0.09178400039672852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,768,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,256,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,64,0.006095110956165526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,32,0.01402844488620758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,32,0.02052622205681271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,1536,0.052674667702780835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,12288,0.028479999966091577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,7168,0.032237334383858576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,2048,0.005571555760171678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,10240,0.16066843933529323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,2048,0.016827555166350473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,16384,0.09392888678444757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,2560,0.010246222217877706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,4096,0.0314319994714525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,51200,3.789340549045139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,10240,0.8998692830403646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,512,0.08979733122719659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,7168,0.04251911242802938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,12288,0.04305244485537211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,768,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,3584,0.021368889345063105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,65536,0.07052710983488295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,512,0.003995555556482739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,6144,0.010575999816258749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,3072,0.02513155506716834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,6144,0.019175110591782462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,1536,0.010204444328943888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,4096,0.009837333526876237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,1536,0.010639111200968424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,8192,0.014815110299322339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,6144,0.015063999427689446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,1536,0.011153777440388998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,4096,0.07480622000164457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,512,0.005606222069925732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,10240,0.04709955718782213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,512,0.034908443689346313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,32,0.006704888823959563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,6144,0.018287110659811232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,256,0.004366222355100843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,256,0.013157332936922709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,12288,0.01847999956872728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,768,0.019127110640207928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,16384,0.04211466511090597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,6144,0.04134488768047757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,128,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,2560,0.16744800408681235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,5120,0.017126222451527912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,1536,0.005610666755172942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,2048,0.007394666473070781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,2048,0.008541333178679148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,4096,0.18703111012776694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,12288,0.11941866079966228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,6144,0.11817244688669841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,3072,0.028143111202451918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,512,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,128,0.008780444661776224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,128,0.005007110950019625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,10240,0.0324853327539232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,1024,0.016922665966881644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,128,0.008788444101810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,512,0.018039999736679923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,12288,0.01959022217326694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,512,0.005459555735190709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,2560,0.021536888347731695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,3072,0.025207999679777358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,10240,0.14926577938927546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,2560,0.012764444781674279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,32,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,6144,0.01239733315176434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,32,0.0046533333758513136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,10240,0.02343822187847561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,5120,0.013150222599506378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,51200,0.07611111137602064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,4096,1.3801262113783095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,10240,0.03947822252909342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,768,0.004515555583768421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,2048,0.013651555611027611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,12288,0.026778666509522334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,4096,0.01240711079703437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,6144,0.012854222622182635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,64,0.02743822170628442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,7168,0.32915645175509983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,4096,0.012502222425407834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,6144,0.1854399972491794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,4096,0.009563555320103964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,10240,3.598528120252821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,3072,0.013532444834709167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,16384,0.07390222284528944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,2048,0.007393777370452881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,128,0.1060791081852383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,5120,0.010189333723651038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,1536,0.006353777729802662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,4096,0.030245333909988403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,2048,0.008781332936551835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,16384,0.2970942126380073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,4096,0.01962577799956004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,2048,0.007012444237867991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,1024,0.009141333401203156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,6144,0.046258668104807533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,2560,0.016716443830066256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,768,0.009089777866999308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,32,0.0074480002125104266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,10240,0.20628533098432753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,51200,1.6421564949883356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,7168,0.03219822380277846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,1536,0.006079111248254776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,768,0.025770665870772466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,2560,0.2120880020989312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,1024,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,2560,0.015558222929636637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,5120,1.0148835712009006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,32,0.008055110772450766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,6144,0.09864533609814113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,3072,0.08093955781724718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,768,0.004855999930037392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,10240,0.019675556156370375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,51200,0.07011466556125216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,12288,0.09832977586322361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,16384,0.03190044562021891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,8192,0.1663395563761393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,65536,0.2261857721540663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,1024,0.004209777961174647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,3584,0.014474666780895658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,2048,0.0075253330998950535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,7168,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,6144,0.9896524217393664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,51200,0.24168888727823892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,7168,0.16301066345638698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,12288,0.04179999894566006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,2048,0.007321777443091075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,1536,0.011060444845093621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,512,0.006986666884687211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,768,0.025452444950739544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,1024,0.34494932492574054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,128,0.0053191110491752625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,512,0.01943733294804891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,256,0.0038408889538711975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,10240,0.018218666315078735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,32,0.006115555349323485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,8192,0.02332533399264018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,2048,0.019824888971116807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,1024,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,1024,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,7168,1.6597617467244465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,8192,4.0033955044216585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,64,0.006823110911581252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,10240,0.16214844915601942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,3072,0.01738755570517646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,4096,0.01718844473361969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,16384,0.39372889200846356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,768,0.00370399985048506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,2048,0.14401955074734157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,8192,2.5181831783718533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,7168,0.07798844575881958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,3584,0.34222666422526044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,512,0.056009776062435575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,256,0.014431110686726041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,12288,0.04632355438338386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,32,0.007297777467303806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,256,0.020868443780475195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,10240,0.03165600034925673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,3072,0.16197066836886936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,768,0.00618488879667388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,3584,0.017112889223628573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,512,0.006962666908899943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,16384,0.16973511377970377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,1024,0.009341333475377824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,6144,0.1974897782007853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,768,0.01109333336353302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,51200,0.16714844438764784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,32,0.0027457777824666765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,51200,0.1408053371641371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,12288,0.017177777157889474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,512,0.006026666611433029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,768,0.0037902221083641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,5120,0.01982399986849891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,128,0.09104800224304199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,7168,0.03297955460018582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,10240,4.087375217013888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,1536,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,512,0.0047084444926844705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,12288,0.09236622518963283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,3584,0.051411555873023145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,128,0.21920443905724418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,64,0.009324444664849175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,128,0.007686222592989604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,6144,0.1315066681967841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,768,0.0189057770702574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,5120,0.08952799770567153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,32,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,65536,11.436830308702257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,2560,0.02251022226280636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,3072,0.0306986669699351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,5120,0.6812666787041558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,6144,0.011424889167149862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,64,0.0038702223036024305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,128,0.006275555739800136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,512,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,8192,0.1519120004442003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,32,0.015392889579137167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,256,0.008958222137557136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,1024,0.006008000009589725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,10240,0.019745777050654095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,768,0.0041200001206662925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,128,0.005286222116814719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,16384,0.02289955483542548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,16384,0.24719378683302137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,2048,0.01612444387541877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,2048,0.05142044358783298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,2048,0.044439110491010875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,65536,0.2728639973534478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,8192,0.021935110290845234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,51200,0.1364319986767239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,2560,0.012876444392734103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,128,0.00850577818022834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,256,0.004350222233268949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,512,0.004456888884305954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,51200,0.15478666623433432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,256,0.031517333454555936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,768,0.22416443294949004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,51200,1.0497173733181424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,16384,0.04550666610399882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,64,0.8987831539577908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,3584,0.09913866387473212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,6144,0.023059555225902136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,12288,0.670737796359592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,16384,0.6963804562886556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,32,0.003660444584157732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,16384,0.17817778057522246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,65536,0.9886915418836805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,4096,0.07087466451856825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,8192,0.09164622094896104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,51200,0.058352887630462646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,65536,0.11341599623362224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,5120,0.20381244023640951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,1536,0.03649866580963135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,64,0.012654222548007965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,768,0.02115733259254032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,6144,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,2560,0.07319022549523248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,64,0.0025475554996066624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,6144,0.013557333085272046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,5120,0.009655999640623728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,2048,0.005399111244413588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,32,0.003245333416594399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,10240,0.5301520029703776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,64,0.012868444124857584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,51200,0.08007466793060303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,65536,9.083114624023438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,5120,0.857061333126492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,8192,0.017571555243598092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,6144,0.012357333468066322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,2560,0.06164355410469902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,768,0.033117334047953285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,768,0.009017777939637503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,2048,0.11765866809421116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,4096,0.04006133476893107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,65536,3.7280739678276906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,3072,0.16392355495029026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,3072,0.02972088919745551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,256,0.011576000187132092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,2048,0.0068604445291890045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,51200,0.2222693231370714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,2560,0.027971555789311726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,10240,0.024950222836600408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,12288,0.010148444109492833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,8192,0.0237128883600235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,2048,0.11276888847351074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,3584,0.010762666662534079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,7168,0.19388088915083143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,3072,0.2205280065536499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,128,0.006613333192136552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,3072,0.2066488928265042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,1024,0.006635555376609166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,8192,0.1452373398674859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,3072,0.01089777797460556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,10240,0.046579556332694165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,5120,0.16478044456905788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,12288,0.023637332850032385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,65536,2.982819451226128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,3072,0.011326221956147088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,256,0.02145599987771776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,8192,0.0697751111454434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,51200,0.24216087659200033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,16384,0.03684711125161912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,5120,0.10646755827797784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,64,0.027256889475716486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,65536,0.03555644551912943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,8192,0.018218666315078735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,65536,0.1246088875664605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,1536,0.018816888332366943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,1536,0.014510222607188754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,1024,0.008412444757090675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,6144,0.04420800010363261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,32,0.00535733335547977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,768,0.02113688819938236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,51200,0.17015111446380615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,1536,0.046536889341142446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,8192,0.04976088802019755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,7168,0.0934186644024319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,512,0.0032808888289663526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,1536,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,16384,0.40933956040276426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,65536,0.7020177841186523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,1024,0.010632000035709806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,65536,0.27338043848673504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,51200,1.6672649383544922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,8192,0.1093297799428304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,64,0.009695111049546136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,2560,0.03363022208213806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,128,0.003914666672547658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,4096,1.1723004447089302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,8192,0.07240800062815349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,16384,0.47502756118774414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,10240,0.15986843903859457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,51200,0.3254088825649685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,1024,0.02420177724626329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,32,0.006623111251327727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,2048,0.009425777528021071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,256,0.0032604444358083936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,3072,0.013034666577974955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,2560,0.006015111174848344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,512,0.006491555521885554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,256,0.02088000045882331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,1536,0.009830222361617619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,512,0.005623999983072281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,128,0.008575110799736446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,128,0.00609422226746877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,5120,0.00612088872326745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,7168,0.030615998639000788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,512,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,7168,0.10542222526338364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,32,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,2560,0.006816888848940532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,10240,0.8854853312174479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,8192,0.21713333659701875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,10240,0.04756711257828606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,51200,1.4432915581597223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,512,0.43745867411295575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,4096,0.8345893224080404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,3584,0.08927910857730442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,64,0.003934222377008862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,2560,0.34939734141031903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,5120,0.02335022224320306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,1024,0.004736888739797804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,16384,0.9835359785291883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,4096,0.030321776866912842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,16384,0.05031555559900072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,65536,0.790014213985867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,7168,0.3122177653842502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,7168,6.013094160291884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,51200,0.09546311034096612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,3584,0.013517333401574029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,32,0.010098666780524785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,128,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,16384,0.11198043823242188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,12288,0.1272346708509657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,16384,0.2887635495927599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,2560,0.01697688963678148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,16384,0.09040533171759711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,6144,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,4096,0.02053333322207133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,5120,0.19507200188106963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,1024,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,32,0.002790222151411904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,6144,0.06251466936535306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,4096,0.25403467814127606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,2560,0.08078222142325507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,4096,0.010009778042634329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,2560,0.03889155719015334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,3072,0.021191999316215515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,3072,0.02176888949341244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,3584,1.175520896911621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,7168,0.03772355450524224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,16384,0.6414542198181152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,7168,0.08414489030838013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,1024,0.00702133360836241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,4096,0.01579466627703773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,768,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,16384,0.1388782262802124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,2048,0.013576000101036496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,3072,0.033779554896884494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,2048,0.008486222061845992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,1536,0.07973866992526583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,768,0.012300444145997366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,12288,0.21336532963646782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,256,0.009945777555306753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,65536,0.13755644692314997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,6144,2.1524196200900607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,6144,0.03479555580351088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,6144,0.20324177212185332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,10240,0.02810844447877672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,256,0.008194666769769456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,1536,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,3584,0.019917334119478863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,32,0.0025724443710512584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,2048,0.07894755734337701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,3072,0.013311111264758639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,64,0.005324444423119227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,51200,0.21391377184126112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,1024,0.005960000058015187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,8192,0.01408622165520986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,2560,0.01495911180973053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,4096,0.0103511114915212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,5120,0.01479022204875946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,128,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,65536,2.738744947645399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,128,0.0038453332251972626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,1536,0.009465778039561378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,4096,0.15215378337436253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,1536,0.014257777068350049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,12288,0.05769244167539808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,512,0.27290132310655385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,6144,0.354294220606486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,51200,0.5697555541992188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,65536,0.7261173460218641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,2048,0.03669955664210849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,1536,0.10721866952048408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,65536,0.15520889229244655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,3584,0.12878578239017063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,10240,0.026661333110597398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,64,0.010823999842007955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,512,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,32,0.009397333694828881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,2560,0.0076453329788313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,4096,0.2285146713256836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,12288,0.022167111436525982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,3072,0.0163137780295478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,12288,0.02310133311483595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,8192,0.017095110482639737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,64,0.006120000034570694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,2048,0.014023999373118082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,65536,0.4309057659573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,7168,0.03377155462900797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,768,0.005208888815508948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,4096,0.008259555531872643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,10240,0.04071911176045736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,5120,0.04981422093179491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,7168,0.023758222659428913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,7168,0.01313688854376475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,2560,0.011681777735551199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,5120,0.01953155630164676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,65536,0.011434666812419891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,1536,0.03159200151761373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,512,0.0052933332820733385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,4096,0.037288000186284385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,3072,0.052397333913379245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,128,0.033165332343843244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,512,0.0034168888297345904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,6144,0.1653697755601671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,51200,0.05415644579463535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,2560,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,5120,0.007240888973077138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,128,0.0043333334227403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,7168,0.017248888810475666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,256,0.003350222276316749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,10240,0.01608088943693373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,2560,0.007636444436179266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,512,0.003629333443111844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,1024,0.007963555554548899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,3072,0.013023110727469126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,65536,0.21871644920772979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,3072,0.029109332296583388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,128,0.004337777694066365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,12288,0.030191110240088567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,512,0.007903111477692923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,3584,0.017122666041056316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,1024,0.07572622431649102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,1536,0.03302399979697333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,768,0.02772177755832672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,7168,0.011368888947698804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,51200,0.19038311640421549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,2560,0.010050666828950247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,12288,0.042176001601748996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,5120,0.013344888885815939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,16384,0.47668176227145725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,2560,0.006027555714050929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,2048,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,7168,0.4271804491678874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,256,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,256,0.006196444647179709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,65536,0.20710221926371256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,4096,0.010873777998818291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,512,0.18698488341437447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,2048,0.006353777729802662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,768,0.0037786666717794207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,768,0.0058764442801475525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,1024,0.01366133325629764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,16384,0.011630222201347351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,8192,0.033831109603246055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,8192,0.01019022199842665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,1024,0.09647199842664932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,2560,0.06098133325576782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,256,0.010764444039927589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,12288,1.79931640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,1024,0.0058222222659322954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,128,0.006646222124497096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,6144,0.028347555134031508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,8192,0.14258755577935114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,12288,0.6023626857333714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,2560,0.016270221935378183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,12288,0.3344577683342828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,16384,0.052968889474868774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,1024,0.01920622256067064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,1536,0.013777777552604675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,5120,0.034854223330815635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,12288,0.02880711025661892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,3584,0.01722666621208191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,65536,0.034134222401512995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,5120,2.5368569691975913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,64,0.009364444348547194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,1024,0.015222221612930298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,2560,0.2902577718098958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,2560,0.04210933380656772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,1536,0.010799999866220685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,7168,0.21908533573150635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,4096,0.01423111061255137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,1536,0.05489066574308607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,51200,0.12600355678134495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,12288,0.03237688872549269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,51200,0.28012887636820477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,65536,6.838541242811416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,2048,0.00829777783817715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,65536,0.09667199850082397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,1024,0.006833777659469181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,1536,0.02126311096880171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,7168,0.013576889203654395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,12288,0.018820444742838543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,4096,0.012585777375433179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,256,0.004592000196377437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,1536,0.016116445263226826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,10240,0.02149155570401086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,8192,0.010833777487277985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,128,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,3072,0.05687555339601305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,4096,0.12206578254699707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,2048,0.01867288847764333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,2560,0.0075911109646161394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,12288,0.10208355718188816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,512,0.025320000118679468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,4096,0.01536622146765391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,7168,0.018011555075645447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,64,0.008620444271299574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,3072,0.010871111518806882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,16384,0.23734132448832193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,128,0.004209777961174647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,2048,0.0058044443527857465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,8192,0.400765339533488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,1536,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,1024,0.037775112522972956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,64,0.0031795555518733132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,128,0.004081777814361784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,51200,0.3478399912516276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,65536,1.9483555687798395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,512,0.00600622221827507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,128,0.004486222234037188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,4096,0.029158221350775823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,51200,0.02663733396265242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,16384,0.13855377833048502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,51200,1.987704806857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,768,0.010463111102581024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,3072,0.016033778587977093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,256,0.006426666759782367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,7168,0.04872622092564901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,51200,0.06337333387798734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,51200,0.6898861991034614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,7168,0.023219555616378784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,7168,0.011731555892361535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,4096,0.09063022004233466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,3584,0.06010044283337063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,4096,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,3584,0.015765332513385348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,65536,0.09407999780442979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,6144,0.024331554770469666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,65536,1.2447306315104167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,7168,0.04758488800790575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,256,0.006241777704821692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,65536,2.692906697591146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,12288,0.023286221755875483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,32,0.010808000134097205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,512,0.004631999880075455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,32,0.00720977783203125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,10240,0.032023111979166664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,32,0.012064889073371887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,65536,0.1412595510482788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,3584,0.011966221862369113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,5120,0.0828088919321696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,65536,0.09146310885747273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,6144,0.05446933375464546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,512,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,12288,0.03611377875010172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,512,0.0034684443639384378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,51200,0.05407199925846524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,3072,0.02327466673321194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,6144,0.021307556165589228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,1024,0.008798222574922774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,3584,0.05327733357747396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,12288,0.3591510984632704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,10240,0.1254862282011244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,51200,1.2263938056098092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,4096,0.02456355591615041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,1024,0.04916355676121182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,6144,0.10411289003160264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,256,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,10240,0.03730310996373495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,2048,0.01588711142539978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,10240,0.045935998360315956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,7168,0.07359555694792005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,65536,0.031104889180925157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,5120,0.02239377796649933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,2048,0.04520888792143928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,7168,0.016858667135238647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,16384,0.023036445180575054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,12288,0.04900266726811727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,6144,0.02979733215437995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,3072,0.01351111133893331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,32,0.007010666860474481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,7168,0.03491200009981791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,8192,0.19575111071268717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,6144,0.024862221545643274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,10240,0.01367288910680347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,65536,0.06587466928693983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,1024,0.004987555659479565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,2560,0.008213332957691616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,768,0.022591999835438196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,128,0.004528000122971005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,16384,0.020794666475719877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,12288,0.07967733012305366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,12288,0.2350320021311442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,128,0.0058871110280354815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,512,0.01867733399073283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,768,0.005932444499598608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,16384,5.784536149766709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,7168,0.11929066975911458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,6144,0.03903733359442817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,4096,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,6144,0.011174221833546957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,256,0.042583111259672374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,2560,0.009111999637550777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,3072,0.017536888519922893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,128,0.06669866376452975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,128,0.013470222552617392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,2560,0.005518222020732031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,6144,0.07231110996670194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,256,0.004634666774008009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,65536,0.17338132858276367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,1536,0.052332444323433765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,2048,1.7751635445488825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,32,0.0026906666656335196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,1536,0.0517484446366628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,5120,0.045532445112864174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,6144,0.020625778370433383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,64,0.0036159998012913596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,128,0.003949333396222857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,65536,0.11269688606262207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,12288,0.3289857705434163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,3584,0.030004445049497817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,16384,3.7028104994032116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,768,0.004991111242108875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,64,0.016559110747443307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,8192,0.06806933217578463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,51200,0.5154533386230469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,256,0.017702221870422363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,65536,0.5242924690246582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,1536,0.04353777898682488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,64,0.0791191127565172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,8192,0.02056533263789283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,1536,0.18036354912651908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,32,0.005127999931573868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,2560,0.006240889016124938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,512,0.005877333382765452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,32,0.004864000197913912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,1536,0.006630222416586346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,6144,0.011509333219793109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,256,0.0041146667467223275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,512,0.005983110931184556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,16384,0.5216933356391059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,32,0.009911999934249455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,1536,0.10575199789471096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,768,0.005950222412745158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,51200,1.6374684439765081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,1536,0.02777155571513706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,3072,0.017267554998397827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,10240,0.05297333333227369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,256,0.01661155621210734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,5120,0.22533867094251844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,10240,0.041402667760849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,65536,2.4324987199571395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,256,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,1536,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,7168,0.12743555174933538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,12288,0.039854221873813205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,8192,0.07203733258777194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,512,0.015969778100649517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,128,0.009611555271678502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,256,0.007084444165229797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,12288,0.02401688860522376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,6144,0.019090667366981506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,12288,0.02070844504568312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,2048,0.007715555528799693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,1536,0.05877688858244154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,7168,0.2535617881351047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,4096,0.02896711230278015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,2560,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,6144,0.1926542255613539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,7168,0.010975999964608086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,1536,0.05237599876191881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,10240,0.02730577852990892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,4096,0.007843555675612556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,3072,0.007376888559924231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,65536,2.0219012366400824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,2048,0.009758222434255812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,768,0.030869334936141968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,6144,0.09297777546776666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,6144,0.1067848867840237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,3584,0.010572444233629437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,2560,0.009986666341622671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,4096,0.02463200026088291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,64,0.005761777775155173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,65536,0.24217600292629668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,768,0.004831111265553368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,8192,0.03492000036769443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,64,0.012345777617560493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,1024,0.0042160000238153665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,10240,0.056272890832689076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,768,0.00853066643079122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,768,0.00888533310757743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,16384,0.0588773356543647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,6144,0.012215110990736218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,128,0.0026986667265494666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,768,0.010837333069907295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,7168,0.05884444713592529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,2048,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,65536,2.300291485256619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,16384,0.1461697816848755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,12288,0.009979555176364051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,51200,0.08810489045249091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,65536,0.3498266537984212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,768,0.012159999873903064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,1536,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,51200,0.20236178239186606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,16384,0.06019555860095554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,6144,4.397013346354167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,51200,2.052492353651259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,12288,0.03086844417783949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,768,0.009483555952707926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,2048,0.032588445478015475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,1536,0.008031110796663497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,1536,0.12511200375027126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,2560,0.016508445143699646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,4096,0.012280888855457306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,8192,0.10200711091359456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,10240,0.039523555172814265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,6144,4.638671027289496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,2560,0.1042684449089898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,3072,0.011666666302416058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,64,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,8192,41.01811048719618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,64,0.008235555556085374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,512,0.009106666677527958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,12288,0.02241866621706221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,5120,0.10890489154391819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,512,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,51200,0.9980675379435221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,1536,0.010851555400424533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,2560,0.014374223020341663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,2560,0.3673546579149034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,16384,0.18812621964348686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,16384,0.023235556152131822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,6144,0.009877333210574256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,4096,0.7805155648125542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,256,0.006107555495368109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,65536,0.2932071155971951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,12288,0.01702844434314304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,65536,0.19520444340176055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,51200,0.023766222927305434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,3584,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,65536,0.321943998336792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,512,0.0047120000753137796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,16384,0.05646222167544895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,10240,0.04566755559709337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,1024,0.02142222225666046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,512,0.005568888866239124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,6144,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,256,0.026727111803160772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,4096,0.009357333183288574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,7168,0.009000889129108852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,2560,0.04328088959058126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,768,0.005331555588377847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,5120,0.024012444747818842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,4096,0.1530124478869968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,12288,0.15884799427456325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,256,0.02015911042690277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,256,0.004586666822433472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,51200,0.2285519970787896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,32,0.00509333320789867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,12288,0.037075555986828275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,3584,0.0247688889503479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,10240,1.8264444139268663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,12288,0.19330755869547525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,16384,0.3577262295616998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,65536,0.06587555673387316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,3584,0.13545600573221842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,2048,0.017489777670966256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,6144,0.02806666824552748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,3584,0.030198223061031763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,768,0.0037751110891501107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,5120,1.4816115697224934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,3584,0.024921778175565932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,10240,0.019052444232834708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,1536,0.0069955554273393415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,65536,0.017309332887331646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,3584,0.010792888700962067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,8192,0.01555288831392924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,4096,0.008223111430803934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,3584,0.040580444865756564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,3584,0.08906400203704834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,2560,0.23376801278856066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,12288,0.03360177742110358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,65536,2.1065154605441623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,6144,0.018030222919252183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,2048,0.44173065821329754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,64,0.052052444881863065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,65536,0.13711288240220812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,4096,0.7459191216362847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,12288,0.92995728386773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,768,0.0467217763264974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,7168,0.2635493278503418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,2560,0.013977777626779346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,32,0.003879111260175705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,7168,0.4340177906884088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,10240,0.08002489142947726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,8192,0.01650311052799225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,12288,0.17987467183007133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,12288,0.047022223472595215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,1024,0.010401777923107147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,16384,0.036496887604395546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,3584,0.01847466660870446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,2048,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,16384,0.052744001150131226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,128,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,32,0.006511999915043513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,5120,0.8871413336859809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,5120,0.016833777228991192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,8192,0.0181040002240075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,6144,0.021565333008766174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,128,0.009380444056457942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,1024,0.004425777743260066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,3584,0.0075333333677715724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,65536,0.1409697797563341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,5120,0.008809777597586313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,5120,0.09364000293943618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,64,0.01662577854262458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,64,0.00508088908261723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,51200,0.07490044169955783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,64,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,65536,0.016952888833151925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,3072,0.02271733350223965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,32,0.0037128888070583344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,1024,0.012902222573757172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,128,0.003721777763631609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,10240,0.9795582029554578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,2560,0.013215999636385175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,12288,0.034733331865734525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,8192,0.02236266599761115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,51200,0.14682043923272028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,6144,0.028246220615175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,8192,0.014547554983033074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,8192,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,12288,0.016421332955360413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,2560,0.013212444053755866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,12288,0.2388239966498481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,10240,0.1083022223578559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,65536,0.04260799951023526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,7168,0.126108447710673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,8192,0.012441777520709567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,12288,1.1419590844048393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,512,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,6144,0.019560888409614563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,51200,0.16465244028303358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,2048,0.009024889104896123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,6144,0.09633155663808186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,12288,0.01590488851070404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,1536,0.020572443803151447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,1024,0.01499644418557485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,2560,0.12785777780744764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,2560,0.028096887800428603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,512,0.05589244763056437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,4096,0.013188444077968597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,256,0.3861733277638753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,4096,0.01462222304609087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,1536,0.08347733153237237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,768,0.011692444483439127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,3072,0.00849333322710461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,256,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,256,0.01052266690466139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,5120,0.02214488883813222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,12288,0.45736180411444766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,8192,0.20151289304097494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,3584,0.3673715591430664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,32,0.006313777632183499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,2560,1.1387181811862521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,2560,0.02125155594613817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,6144,0.024123556084103052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,1536,0.0163102216190762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,3072,0.011991110940774282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,256,0.006135111053784688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,1024,0.015099555253982544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,3584,0.0639537771542867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,3072,0.011705777711338468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,8192,0.1998293267356025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,4096,0.03938400083118015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,1536,0.3886257807413737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,8192,0.05238577723503113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,6144,0.2083991103702121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,512,0.03369066781467862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,1024,0.012053333222866058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,10240,0.08732622199588352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,3072,0.026709333062171936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,64,0.011213333242469363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,1536,0.014951111541854011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,51200,0.16993244489034018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,256,0.015559110376569959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,1536,0.006895111252864202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,4096,0.008933333059151968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,128,0.00272088891102208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,12288,1.8101832071940105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,16384,0.028749333487616643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,10240,0.04983377787801954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,7168,0.019492444064882066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,32,0.004313777718279096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,3584,0.009741333623727163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,512,0.007173333730962541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,6144,0.16185688972473145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,5120,0.01016711112525728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,8192,0.012560000022252401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,12288,0.05756711297565036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,12288,0.028208000792397395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,32,0.0029004443850782183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,12288,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,512,0.01219555570019616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,7168,0.47772354549831814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,3584,0.35922577646043563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,3584,0.021049777666727703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,3584,0.5545760260687934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,128,0.005695111221737332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,10240,1.2894684473673503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,3584,0.015490666031837463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,2048,0.013370667066839008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,256,0.004247111164861255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,1024,0.1352062225341797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,3584,0.024001777172088623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,2048,0.06899733013576932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,10240,0.027264000640975103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,2560,0.01998844411638048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,4096,0.010965333216720156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,6144,0.21004977491166857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,6144,0.25419643190171987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,768,0.021946666969193354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,128,0.007541333635648091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,256,0.030167112747828167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,512,0.015846222639083862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,16384,0.41327643394470215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,12288,0.03220444586541917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,3072,0.00830844458606508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,2560,0.03241777751180861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,16384,0.07737955782148573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,2560,0.052537779013315834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,7168,0.013815999858909182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,51200,0.04413333204057482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,768,0.003671110918124517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,64,0.018361777067184448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,51200,2.7854879167344837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,3072,0.09288355376985337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,2560,0.00758844448460473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,128,0.006623111251327727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,8192,0.266881783803304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,128,0.014874666929244995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,128,0.022447110878096685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,8192,0.124309327867296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,5120,0.009334222310119206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,3072,0.25854934586419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,3072,0.008616888688670265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,3584,0.018967999352349173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,1024,0.006223111102978389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,10240,0.030223998758527968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,2560,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,64,0.14398222499423557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,51200,0.055753780735863574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,8192,0.8915981716579862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,51200,19.88241916232639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,5120,0.09440088934368557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,51200,0.1285626623365614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,10240,0.1621288855870565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,64,0.002891555635465516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,6144,0.0451182226339976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,10240,0.3087422317928738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,2560,0.00756444450881746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,65536,0.35707465807596844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,8192,0.20260800255669487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,2560,0.03874222106403775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,4096,0.066392891936832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,768,0.007643555601437886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,16384,0.025422222084469263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,5120,0.2339928944905599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,8192,0.22671021355523002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,4096,0.008472888833946651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,12288,0.39169955253601074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,128,0.07767377959357367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,4096,0.09340177641974555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,1024,0.006313777632183499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,1536,0.03663555449909634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,32,0.00829777783817715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,7168,0.032440887557135686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,1536,0.005640000104904175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,512,0.005762666463851929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,5120,0.10284977489047581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,8192,0.018213333355055917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,12288,0.008485332959228093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,1024,0.005596444424655702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,7168,0.013898667362001208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,2560,0.014955555399258932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,128,0.0037466668420367767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,2560,0.09382400247785781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,2048,0.007173333730962541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,10240,0.13439288404252794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,512,0.0047644442982143825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,10240,0.007703999678293864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,1536,0.004379555583000183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,3584,0.014736000034544202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,3072,0.00871377769443724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,3072,0.04548177785343594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,6144,0.05302844444910685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,7168,0.07374933030870226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,256,0.029211554262373183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,65536,0.059296886126200356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,1024,0.01353422221210268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,3072,0.008103110724025303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,256,0.003116444374124209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,32,0.005151999907361137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,8192,0.5441057946946886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,2048,0.04330933425161573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,768,0.011606222225560082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,1536,0.023353776997990076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,6144,0.013208000196350945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,8192,0.3559929000006781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,1536,0.015475556254386902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,2560,0.00905866672595342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,7168,0.015857777661747403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,128,0.008632889224423302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,512,0.01018044435315662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,128,0.010149333212110732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,3584,0.016406221522225272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,16384,0.03965511255794101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,3584,0.7749919891357422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,6144,0.08441244231330024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,5120,0.00941777808798684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,3584,0.04623200164900886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,32,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,16384,0.05840888950559828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,64,0.003563555578390757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,7168,0.027739556299315557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,128,0.2972808943854438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,2048,0.013029333617952136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,768,0.02514755560292138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,2048,0.007487110793590546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,32,0.0026782221264309357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,1024,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,4096,0.021464889248212177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,32,0.004498666773239772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,2048,0.025071110990312364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,2048,0.19099733564588758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,3584,0.014363555444611443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,768,0.008261333737108443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,512,0.10491022136476304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,4096,0.037343111303117536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,7168,0.05019377668698629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,7168,0.21851466761695015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,3072,0.22482665379842123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,3072,0.016542222764756944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,16384,0.015188443991872998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,8192,0.6019920243157281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,7168,0.3576524522569444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,7168,0.019575999842749704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,7168,0.019095111224386428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,128,0.009224889179070791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,256,0.003567999849716822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,51200,2.6988177829318576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,256,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,6144,0.04147911071777344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,3072,0.04126222266091241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,16384,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,256,0.008433777425024245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,65536,4.924779680040148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,512,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,16384,0.04366222355100843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,65536,2.468607160780165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,1536,0.0635475582546658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,512,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,32,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,65536,0.12018667327033149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,32,0.07125511434343126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,768,0.027151111099455092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,256,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,51200,3.5308248731825085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,8192,0.25639110141330296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,7168,0.011935110721323224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,1024,0.008662222160233391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,3584,0.017892445127169292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,7168,0.018377777602937486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,5120,0.017881777551439073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,10240,0.07580088906817965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,64,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,768,0.0377146667904324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,256,0.023822221491071913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,2560,0.0650328861342536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,10240,0.04463644491301643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,1536,0.13429510593414307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,12288,0.04007022248374091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,1024,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,128,0.0028008888992998335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,16384,0.7691084543863932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,256,0.005084444251325396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,65536,0.558892461988661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,51200,0.0766542222764757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,32,0.009746666583749983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,51200,13.060827467176649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,512,0.03337688909636603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,5120,0.11292978127797444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,16384,0.6617893113030328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,5120,0.12103733751508926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,256,0.011602666642930774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,10240,0.15615999698638916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,4096,0.01093066649304496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,768,0.026592888765864905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,4096,0.02424088948302799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,512,0.01166133334239324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,256,0.011714666253990598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,6144,2.3152328067355685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,4096,0.012705778082211813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,128,0.12285689512888591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,4096,0.017851556340853374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,6144,1.8071253034803603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,51200,0.33307909965515137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,2560,0.7117404407925076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,6144,0.06703466839260526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,2048,0.013571555415789286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,2048,0.01276533305644989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,32,0.008443555898136562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,768,0.0037466668420367767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,256,0.01936444474591149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,64,0.011876444849703046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,3072,0.03704977697796292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,2048,0.029301332102881536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,64,0.015154666370815702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,128,0.005920889063013925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,6144,0.01016444464524587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,7168,0.4479546546936035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,768,0.030795554320017498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,64,0.0038408889538711975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,3072,0.027775110469924078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,6144,0.024821332759327356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,32,0.028039998478359644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,1536,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,5120,0.36720800399780273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,10240,0.2659093274010552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,4096,0.31802845001220703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,32,0.006693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,6144,0.0332951115237342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,16384,0.09605955415301853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,768,0.008006222546100616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,4096,0.010346666806273991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,10240,0.08032888836330838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,512,0.009444444543785518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,3072,0.03143466512362162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,3584,0.2837875684102376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,12288,0.2624533441331652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,6144,0.01689511040846507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,128,0.004705777598751916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,10240,0.020763556162516277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,3584,0.7786355548434787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,51200,3.875962575276693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,8192,0.02293955617480808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,2048,0.46927467981974286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,4096,0.010667555862002902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,10240,0.015587555037604438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,768,0.030014223522610132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,2048,0.027466666367318895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,3584,0.018239999810854595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,51200,0.12139644887712266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,64,0.002855111120475663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,4096,0.008960888617568547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,2560,0.00683822234471639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,256,0.004577777865860197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,512,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,12288,0.03570400012864007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,4096,0.011676444775528379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,65536,18.335500081380207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,128,0.005705777969625261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,512,0.028262221150928076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,7168,0.01423288881778717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,256,0.010314666562610203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,768,0.01034577770365609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,512,0.008640000389681922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,16384,0.4521582391526964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,2048,0.010611555642551847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,2048,0.017260443833139207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,768,0.008948444492287105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,16384,0.006122666514582104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,10240,0.29716266526116264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,2048,0.07169777817196317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,512,0.0034204444123639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,3584,0.008240000241332585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,128,0.00433955548538102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,128,0.027662222584088642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,1536,0.02138311167558034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,128,0.004385777645640903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,3584,0.16582400268978542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,65536,0.03081866767671373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,2048,0.008982222113344405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,7168,0.00958933350112703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,768,0.007360888852013483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,128,0.005245333330498801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,256,0.0035502223504914176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,768,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,12288,0.07980444696214464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,12288,0.046485334634780884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,5120,0.05165955424308777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,32,0.022707555029127333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,1536,0.05121422145101759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,3584,0.009754666851626502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,3584,0.05439022183418274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,3584,0.009791110952695211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,512,0.016418667303191293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,128,0.015769778026474845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,256,0.0047102222839991255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,1024,0.029342220889197454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,128,0.04858577913708157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,1024,0.010818666881985135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,768,0.005388444496525659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,3584,0.07489777935875787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,2048,0.006544000158707301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,128,0.03574666712019179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,3072,0.985179583231608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,3584,0.19408267074161104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,2560,0.12613688574896917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,3072,0.009163555171754625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,2048,0.0127688886390792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,4096,0.020703110429975722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,6144,0.236753781636556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,7168,0.0058755555914507965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,4096,0.10843200153774685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,2048,0.024970667229758367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,2048,0.009712888962692684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,768,0.00425866660144594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,7168,0.04940355486339993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,8192,0.24235110812717012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,3584,0.05604800250795153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,3072,0.028871999846564397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,1024,0.023013333479563396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,6144,0.01311466677321328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,4096,0.04318755533960131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,10240,0.017066667477289837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,256,0.005386666705211003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,8192,0.016216889023780823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,3072,0.008248888783984715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,64,0.00721244431204266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,5120,0.010308444499969482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,4096,0.49947288301255965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,5120,0.020410666863123577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,2560,0.07375822464625041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,32,0.013403555585278405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,3072,0.019727999965349834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,8192,0.02179733415444692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,51200,0.7767653465270996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,3072,0.010806221928861407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,2560,0.10608444611231486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,8192,0.0063484443558586975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,4096,0.10527200169033474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,8192,0.05321333474583096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,512,0.007256000406212277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,10240,0.10156266556845771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,5120,0.03829066620932685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,128,0.008014221986134848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,32,0.002546666603949335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,6144,0.016594666573736403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,4096,0.18608178032769096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,3072,0.021993777818149988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,51200,0.14687821600172254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,4096,0.010315555665228102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,2048,0.011950222154458364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,32,0.13791733317905003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,1024,0.00556355549229516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,8192,512,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,64,0.003044444446762403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,512,0.2716764344109429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,64,0.00452533322903845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,256,0.006430222342411677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,1536,0.015842666228612263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,768,0.07444622119267781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,4096,0.07830133040746053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,1024,0.09129866626527573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,1536,0.010135999984211391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,51200,0.9005075030856662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,1536,0.00833777752187517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,128,0.16440178288353816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,6144,0.011883555187119378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,3584,0.010816000401973724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,512,0.00500177757607566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,768,0.02182933357026842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,5120,0.026975999275843304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,10240,0.05195644497871399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,8192,0.031758222315046526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,3584,0.9138480292426215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,512,0.013843555417325763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,512,0.007831110722488826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,16384,0.053027557002173535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,4096,0.018966222802797954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,128,0.009418666362762451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,1024,0.09759289026260376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,16384,0.049154665735032826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,3584,0.025153777665562097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,512,0.005557333429654439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,3072,0.13549155659145778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,6144,0.035028444396124944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,128,0.005594666633341048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,51200,1.8913404676649306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,65536,0.22965423266092935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,65536,0.07096711132261488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,2560,0.007354666789372762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,12288,0.13606044981214735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,3584,0.007155555817815993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,7168,0.15683910581800672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,3072,0.010844444235165914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,8192,0.016703999704784818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,6144,0.019009777241282992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,3584,0.07705422242482503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,8192,0.027277333868874445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,256,0.011301333705584208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,768,0.005040888984998067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,7168,0.010988444089889526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,768,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,768,0.00629333323902554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,1536,0.01202666676706738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,1024,0.011163555913501315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,256,0.08632977803548177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,128,0.0086986662613021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,2048,0.047797332207361855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,256,0.01800355480776893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,768,0.005495111147562663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,1536,0.005888000130653381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,5120,0.44785600238376194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,2560,0.025476444098684523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,6144,0.01316088851955202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,12288,0.02442755632930332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,128,0.002971555623743269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,51200,0.10623644457923041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,2560,0.08793777889675564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,768,0.0539244446489546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,7168,0.025014221668243408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,1536,0.037464890215131975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,64,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,256,0.05421511001057095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,12288,0.30192356639438206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,3072,0.009193778038024902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,2560,0.06086133586035835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,32,0.004962666581074397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,3072,0.24158755938212076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,12288,0.03703555464744568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,3584,0.012872888810104795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,8192,0.027317333552572463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,512,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,512,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,128,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,5120,0.23920978440178764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,8192,0.023898666103680927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,128,0.0035013332962989807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,51200,0.058020446035597056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,3584,0.00794400026400884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,768,0.00424888895617591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,3584,0.015600000818570455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,3584,0.00810933361450831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,8192,1.4271556006537542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,512,0.0058551111982928375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,7168,0.12594221697913274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,128,0.015200000670221118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,8192,0.05116888880729675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,1536,0.07759555843141344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,128,0.007401777638329401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,64,0.010651555326249866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,10240,0.021104888783560857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,1024,0.010042666561073726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,4096,0.6167679892645942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,768,0.014576888746685453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,7168,0.03172444303830465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,256,0.004338666796684265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,1536,0.041947556866539844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,10240,0.02281066609753503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,2048,0.02801777919133504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,128,0.004560889055331548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,4096,0.13190666834513345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,51200,1.355913798014323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,16384,0.07263111405902438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,12288,3.246206283569336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,51200,0.11284444067213271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,32,0.0032008888406885993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,2048,0.22933689753214517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,16384,1.744948493109809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,64,0.0041680000722408295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,256,0.00479288895924886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,4096,0.5953991148206923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,4096,0.03644533289803399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,7168,0.04191022117932638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,7168,0.6833715438842773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,5120,0.016720889343155753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,7168,0.03773155477311876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,5120,0.28127911355760365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,5120,0.018001778258217704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,6144,0.09429244200388591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,1024,0.015776000089115567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,16384,0.021474666065639917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,8192,0.08878755569458008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,2560,0.8858933448791504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,3072,0.04725066820780436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,256,0.0169840008020401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,768,0.005635555419656966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,7168,0.07169066535101996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,1536,0.11605600516001384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,256,0.013849777479966482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,2560,0.013252444565296173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,128,0.006723555425802867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,10240,0.32821689711676705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,1024,0.17215111520555285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,6144,0.027609777119424608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,128,0.007377777662542131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,32,0.016718222035302054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,2560,0.012612444659074148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,3584,0.02109866672092014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,3584,0.8639715512593588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,2048,0.015091554986106025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,5120,0.023379556006855432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,2560,0.008628444539176093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,12288,0.006917333437336816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,512,0.008595555192894405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,7168,0.2452871004740397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,7168,0.010771555205186209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,256,0.024325332707828943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,65536,0.22066133552127412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,16384,8.468615214029947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,12288,0.00902844468752543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,16384,0.006160888820886612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,2560,0.007956444389290279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,8192,0.046238223711649575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,3584,0.014721777704026965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,64,0.00507377791735861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,32,0.005174222091833751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,51200,32.10562472873264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,6144,0.014274666706720987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,1024,0.005271111097600725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,8192,0.028335111008750066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,6144,0.6425306532118056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,256,0.073253333568573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,2048,0.005917333480384614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,2560,0.01202044470442666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,128,0.011819555527634092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,1024,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,128,0.013117333253224691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,1024,0.003928889003064898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,6144,0.2775697708129883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,2560,0.012255999777052136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,10240,0.11109066671795315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,10240,0.7714195781283908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,7168,0.23103645112779406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,768,0.00598933340774642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,2048,0.006774222271309958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,128,0.006177777631415262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,4096,0.014051554931534661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,4096,0.19462400012546113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,3584,0.10545244481828477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,2560,0.009393778112199571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,512,0.00589155571328269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,6144,0.03289777702755398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,65536,1.210824860466851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,768,0.012240889171759287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,12288,0.5060453414916992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,3072,0.021013332737816706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,3584,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,32,0.005445333404673471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,32,0.006504888749784893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,2560,0.01223288890388277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,3584,0.2626826763153076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,12288,0.020358221398459542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,256,0.0046746668716271715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,10240,0.028765334023369685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,512,0.007737778127193451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,12288,0.06874844763014051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,7168,0.177894221411811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,3584,0.03126488791571723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,512,0.004823110997676849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,512,0.007620444728268518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,3072,0.025489777326583862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,64,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,4096,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,768,0.033322665426466204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,12288,0.05035466618008084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,51200,0.3299395508236355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,10240,0.013516444298956128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,65536,1.4040560192532008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,3584,0.09613688786824544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,5120,0.02234755622016059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,8192,0.0772542224989997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,1024,0.08396533462736343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,65536,0.4009546703762478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,2048,0.010434666441546546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,6144,0.08885777658886379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,1024,0.008893333375453949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,65536,0.06518933508131239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,64,0.023119999302758112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,512,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,3584,0.07339644432067871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,768,0.008169777691364288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,12288,0.026371555195914373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,3072,0.007724444071451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,1024,0.009186666872766282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,10240,0.19125866889953613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,6144,0.07339021894666883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,64,0.005552000055710475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,128,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,32,0.003626666549179289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,1536,0.03014577759636773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,3072,0.013584888643688627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,4096,0.13743466801113552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,1024,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,256,0.019742222295867074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,8192,0.9228027131822375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,3072,0.15235555171966553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,8192,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,8192,0.1572657823562622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,4096,0.02769600020514594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,65536,1.637139532301161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,2048,0.3192933400472005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,3072,0.10589777761035496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,1024,0.012680889003806643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,3584,0.3054853280385335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,256,0.007803555164072249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,1024,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,512,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,3072,0.007138667007287343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,768,0.022755554980701868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,128,0.028324445088704426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,10240,0.44358134269714355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,7168,0.00666222224632899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,5120,0.020394666327370536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,768,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,1536,0.005693333430422678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,2048,0.03513777918285794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,1536,0.034338666333092585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,65536,0.0680640008714464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,1536,0.0079644446571668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,3584,0.012758221891191272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,2048,0.013391999734772576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,2048,0.038013332419925265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,10240,0.017677333619859483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,128,0.008585777547624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,1024,0.0057795556883017225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,2560,0.013468444347381592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,4096,0.034437331888410784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,65536,0.29997865358988446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,6144,0.011690666278203329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,1536,0.009694221946928237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,2560,0.0074311114019817775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,768,0.004977777600288391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,3584,0.012847999731699625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,2048,0.007350222104125553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,3584,0.027855998939938013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,4096,0.4023253387875027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,768,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,3584,0.09884533617231582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,768,0.014785778191354541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,3584,0.009549332989586724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,6144,0.009649777577983009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,512,0.029630220598644678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,768,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,16384,0.400818665822347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,256,0.0035031110876136353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,768,0.004635555462704765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,2560,0.05168266759978401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,4096,0.037971556186676025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,1024,0.20843644936879477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,51200,0.09816266430748834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,4096,0.05519466598828634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,64,0.0041946665280395085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,128,0.004259555704063839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,512,0.019654222660594516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,256,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,1536,0.006182222316662471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,51200,0.04341155621740553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,512,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,5120,0.28844088978237575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,6144,0.014501333236694336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,2560,0.04759022262361315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,1024,0.1492284403906928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,51200,1.0994577407836914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,1024,0.013372444444232516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,6144,0.11396800147162543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,2048,0.010285333626800114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,128,0.005751111027267244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,51200,0.4718204604254828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,65536,0.09528088569641113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,65536,1.0677964952256944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,4096,0.07744977871576945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,3072,0.009076444639099969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,128,0.004044444610675176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,12288,0.03146933184729682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,3072,0.025621333056026038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,12288,0.26052623324924046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,768,0.007847111258241866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,512,0.011507555842399597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,5120,3.231871075100369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,4096,0.013125333521101209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,512,0.040935112370385066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,8192,0.5923946698506674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,2560,0.022304889228608873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,128,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,1536,0.006292444550328785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,65536,0.1907564401626587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,8192,0.16661243968539768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,1536,0.011751111182901593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,256,0.004579555657174853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,51200,0.7137377527025012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,3072,0.02407022151682112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,32,0.01754488878779941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,8192,0.01238222254647149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,512,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,3072,0.006445333361625671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,16384,0.27183466487460667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,1536,0.004654222064548069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,7168,0.08790310886171128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,12288,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,6144,0.025050666597154405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,4096,0.033977776765823364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,10240,0.8887777858310275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,4096,0.07143821981218126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,5120,0.13343377908070883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,1024,0.014275555809338888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,4096,0.1426008939743042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,1024,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,65536,0.10903377665413751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,512,0.02244088881545597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,3584,0.00795911086930169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,256,0.006665777828958299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,5120,0.025524444050259058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,1536,0.02281422250800663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,32,0.15031111240386963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,3072,0.014736889137162102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,16384,0.014469333820872836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,5120,0.11802400483025445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,768,0.05995021926032173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,8192,0.007803555164072249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,2048,0.04099199838108487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,7168,0.011609777808189392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,51200,0.11441777812110053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,51200,1.5919617546929254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,64,0.0031999999450312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,768,0.010647111468844943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,128,0.009409777820110321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,6144,0.04138311081462436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,2560,0.38921509848700625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,128,0.005028444445795483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,4096,0.023200000325838726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,4096,0.03853066762288412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,2048,0.217668440606859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,16384,0.27453245057000053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,65536,3.1289884779188366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,51200,0.1712391111585829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,3584,0.012123555772834353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,3584,0.04986222253905403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,2048,0.03188622328970168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,768,0.04517066809866163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,16384,0.19623911380767822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,10240,0.8388729095458984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,16384,0.07043200069003634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,64,0.008157333566082848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,8192,0.2898453341590034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,1536,0.006737777756320105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,32,0.010119111173682744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,768,0.019881778293185763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,6144,0.08957778082953559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,3584,0.02274488906065623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,4096,0.08453155226177639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,3584,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,16384,0.2809084521399604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,8192,0.21116533544328478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,2048,0.0052737775776121355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,6144,0.011752888560295105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,64,0.005168888717889786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,3072,0.24285422431098092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,768,0.004688888788223267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,8192,0.01684799955950843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,1024,0.06461866696675618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,768,0.02381422287887997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,1536,0.005210666606823604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,51200,0.5371040238274468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,32,0.044928001032935254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,8192,0.23097332318623862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,51200,0.05787110990948147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,3072,0.017255110873116385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,1024,0.009675555759006077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,51200,0.048384000857671104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,1024,0.004074666649103165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,4096,0.011040888726711273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,12288,1.451987584431966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,64,0.03399644295374552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,1024,0.03454577922821045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,10240,0.18610310554504395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,768,0.004616888860861461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,10240,0.04809777935345968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,6144,0.019089778264363606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,7168,0.056254221333397754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,512,0.0038284444146686127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,1536,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,256,0.01940000057220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,3072,0.396428452597724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,32,0.008593777815500895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,7168,0.012294222083356647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,10240,0.04082311193148295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,64,0.0029208889851967492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,6144,0.020767110917303298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,8192,0.019263111882739596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,3584,0.006383111079533895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,3072,0.0779342254002889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,1536,0.02962933315171136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,6144,0.010565333068370819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,8192,0.02128444446457757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,128,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,256,0.00490133340160052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,65536,0.07468177874883015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,10240,0.28381776809692383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,32,0.0026168889469570587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,51200,2.5331039428710938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,256,0.0069191112286514705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,4096,0.012266666524940066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,10240,1.8601947360568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,51200,1.339654180738661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,6144,0.11801689200931126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,64,0.0031555555760860443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,3584,0.18142667081620958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,64,0.005582222094138463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,1536,0.010466666685210334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,128,0.007878222399287755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,128,0.02463200026088291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,51200,1.6685066223144531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,16384,0.5896426836649576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,65536,0.8665733337402344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,256,0.0041200001206662925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,768,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,5120,0.009292444421185387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,3584,1.6462649239434137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,768,0.006144888699054718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,16384,0.40987200207180446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,51200,0.2436951001485189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,1024,0.056997332308027476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,256,0.014899555179807874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,512,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,1024,0.004734222259786394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,64,0.0030942221896515954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,4096,0.03738666574160258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,256,0.006801777829726537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,3584,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,3584,0.00832266691658232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,768,0.005782222168313132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,3072,0.027905778752432928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,3584,0.09551288684209187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,2560,0.02455644475089179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,3584,0.007163555257850223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,32,0.008975110948085785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,2560,0.015736889508035447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,2048,0.015076445208655464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,10240,0.01751733322938283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,64,0.00301155551440186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,3584,0.07564000288645427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,768,0.0169840008020401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,768,0.164882673157586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,32,0.0038275557259718576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,6144,0.1088355580965678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,16384,0.8446062405904134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,65536,0.21978667047288683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,768,0.011208888557222156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,32,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,4096,0.006205333189831839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,512,0.018944889307022095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,256,0.09537511401706272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,3584,0.030766222212049697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,1536,0.38272799385918516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,3584,0.007743111087216272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,512,0.006020444548792309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,5120,0.03899733225504557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,2560,0.009176889227496253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,8192,0.24884266323513457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,1024,0.018146667215559218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,10240,0.12873600588904485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,512,0.004941333499219683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,51200,0.15991022851732042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,1024,0.0780293345451355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,3584,0.012538666526476542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,2048,0.013280889226330651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,12288,0.2168053388595581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,4096,0.007896888587209914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,128,0.03592533204290602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,8192,0.8873644404941134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,128,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,10240,0.024289778537220422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,12288,0.08031200038062201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,4096,0.1119004487991333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,1536,0.0076453329788313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,1536,0.309242672390408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,2048,0.08688444561428493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,51200,0.1773822175131904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,3072,0.006597333484225803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,16384,0.4248266749911838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,128,0.009136888715955947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,12288,0.02915289004643758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,768,0.003867555409669876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,256,0.004536889079544279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,512,0.005951111101441913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,16384,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,51200,0.08290844493442111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,8192,0.032836443848080106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,2048,0.025587555434968736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,3584,0.01740533279048072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,12288,0.04075110952059428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,65536,32.97449747721354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,3584,0.13885955015818277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,1536,0.009689778089523315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,51200,0.10524621937010024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,65536,0.4979857868618435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,512,0.0033324443631701996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,2560,0.07839555872811212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,3072,0.021056888831986323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,1536,0.06112888786527845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,1536,0.03559022148450216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,1024,0.004051555362012652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,1536,0.00480355570713679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,7168,0.06692622105280559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,512,0.08530844582451715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,8192,0.012051555845472546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,10240,0.03467822074890137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,2048,0.32694845729404026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,512,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,7168,0.402054230372111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,256,0.20350755585564506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,256,0.10220977995130752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,1024,0.032807111740112305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,8192,0.03880533244874742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,128,0.017070222232076857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,1024,0.06005777915318807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,32,0.0352168877919515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,12288,0.041815112034479775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,3584,0.020639111598332722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,1536,0.06276710828145345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,1024,0.005990222096443176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,2560,0.014159111513031853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,10240,0.017509332961506315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,5120,0.041800889703962535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,51200,0.08808355861239964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,16384,0.009706666900051964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,2560,0.00815822184085846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,128,0.005138666679461797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,8192,0.014536889062987434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,3072,0.033358222908443876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,128,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,768,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,3072,0.03493422269821167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,3072,0.15865866343180338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,65536,0.2609662214914958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,5120,0.01089244418674045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,12288,0.030413332912656996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,12288,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,5120,0.20014399952358672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,7168,0.017847110827763874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,8192,0.4106284512413873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,10240,0.062162664201524526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,2048,0.06398666567272611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,768,0.06128622425927056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,4096,0.0058346668051348785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,3584,0.008745777938101027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,768,0.00408000002304713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,16384,0.0855333341492547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,32,0.0033520000676314035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,5120,0.08751822180218166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,1536,0.004962666581074397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,2048,0.007321777443091075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,1536,0.009096889032257928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,256,0.1341866652170817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,6144,0.027162666122118633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,256,0.0070426662762959795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,10240,0.02254666719171736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,64,0.002603555512097147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,12288,0.07446310917536418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,256,0.03300266795688205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,32,0.0033804445217053094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,2560,0.007653333246707916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,1536,0.004688888788223267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,64,0.023738667368888855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,16384,0.21366400188869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,51200,1.1298293007744682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,2048,0.0071928890215026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,256,0.0034479999707804788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,6144,0.04415200153986613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,2048,0.648823102315267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,3072,0.013403555585278405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,2048,0.01937955617904663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,16384,0.09427199761072795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,32,0.011804444094498953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,2560,0.017588444881969027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,6144,0.02647466626432207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,256,0.026869333452648584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,128,0.007304888632562425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,5120,0.08886844582027859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,64,0.02336888843112522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,1024,0.4023742145962185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,32,0.01423822177780999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,1024,0.010743111371994019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,128,0.0047271110945277745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,51200,0.1250844399134318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,1024,0.02662311163213518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,64,0.003129777808984121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,768,0.008936888641781276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,3072,0.007984889050324758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,8192,0.24962133831448027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,256,0.003930666794379552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,256,0.005548444473081165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,8192,0.011705777711338468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,4096,0.008949333594905006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,65536,0.07656000057856242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,3584,0.010080888867378235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,4096,0.026429333620601233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,3072,0.012967999610635968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,512,0.027744000156720478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,10240,0.018351111147138808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,5120,0.016159999701711867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,2048,0.034860445393456355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,10240,0.03342577815055847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,6144,0.129439115524292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,7168,0.2986480130089654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,2560,0.013441777891582914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,5120,0.7660809093051486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,128,0.011508444117175208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,65536,2.903209686279297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,1024,0.016217778126398723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,16384,0.03695199886957804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,128,0.008066666622956594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,1536,0.029263112280103896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,10240,0.02134399943881565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,51200,2.518305672539605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,16384,0.06458933485878839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,6144,0.018283555905024212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,3072,0.08667111396789551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,4096,0.7816942003038195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,768,3.1764365302191844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,6144,1.6969644758436415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,16384,0.06517689095603095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,1536,0.10687910848193699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,10240,0.6381662156846788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,8192,0.020796444680955675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,1536,0.22800887955559623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,10240,0.018263111511866253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,1536,0.5902569029066298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,51200,0.03328711125585768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,7168,0.027798222170935735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,12288,0.07895021968417697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,256,0.004121777911980947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,64,0.003277333246337043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,3584,0.1558871136771308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,65536,0.15157332685258654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,128,0.0047884442740016514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,8192,3.9573919508192272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,5120,18.732833862304688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,10240,0.025032889511850145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,65536,0.2798977692921956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,5120,0.3663946787516276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,10240,1.1441013548109267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,16384,0.7325698004828559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,768,0.057020445664723717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,65536,0.49142758051554364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,5120,0.07754222551981609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,8192,0.018976888722843595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,1536,0.020815999971495736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,12288,0.10891822311613296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,7168,0.024147555232048035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,512,0.006339555399285422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,16384,0.09715822007921006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,128,0.004753777964247597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,2560,0.00832177781396442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,4096,0.05236533284187317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,768,0.015138667490747241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,512,0.006080889039569431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,6144,31.645368787977432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,6144,0.015071999695565967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,3072,0.06960178083843656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,2048,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,256,0.007932444413503012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,3584,0.016548444827397663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,1536,0.006767999794748094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,128,0.01513066722287072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,8192,0.08154044548670451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,3072,0.05549955699178907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,12288,0.8109697765774198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,8192,0.19049155712127686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,5120,0.008882666627566019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,2048,0.010032888915803697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,768,0.004574222283230888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,5120,0.13238399558597141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,2048,0.09862133529451157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,6144,0.02458311120669047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,64,0.010757333702511258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,10240,0.41256266170077854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,3072,0.008632000121805403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,7168,0.05012889040840996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,7168,0.32966934310065377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,1024,0.3354159990946452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,2048,0.008347555167145198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,2048,0.016347555650605094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,512,0.011848889291286469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,768,0.08902844455507065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,10240,0.14909511142306858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,3584,0.14106310738457575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,6144,0.013242666920026144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,3072,0.008036444584528605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,1536,0.03371822171741062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,51200,0.1438408825132582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,16384,0.06751466459698148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,32,0.00286666676402092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,12288,0.2357226742638482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,2560,0.010382222632567087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,7168,0.09584266609615749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,6144,0.031170666217803955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,32,0.0024826667375034755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,2048,0.007116444408893585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,2560,0.015947555502255756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,256,0.07933421929677327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,8192,0.266029331419203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,51200,23.766720241970486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,128,0.008362666600280339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,3072,0.0163111107216941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,65536,0.19259111086527506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,512,0.016919111212094624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,768,0.004907555464241239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,3072,0.006530666516886816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,1536,0.005753777921199799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,256,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,768,0.006637333167923822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,6144,0.03375555409325494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,5120,0.006879111131032308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,51200,5.587864769829644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,32,0.0466195543607076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,51200,0.23139378759596083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,51200,0.08904177612728542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,1024,0.008907555705971187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,3072,0.009106666677527958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,12288,0.015885333220163982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,65536,0.17085333665211996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,5120,0.23245154486762154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,1536,0.006481777876615524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,512,0.005581333405441708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,7168,0.10937866899702285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,768,0.006135111053784688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,10240,0.6951057646009656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,6144,0.013845333622561561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,256,0.03456622362136841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,5120,0.33401423030429417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,8192,0.012980444563759698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,8192,0.015399111641777886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,768,0.020263110597928364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,6144,0.010514666636784872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,12288,0.9107608795166016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,32,0.006771555377377405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,6144,0.055051555236180626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,3072,0.010869333313571082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,8192,0.018927999668651156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,1024,0.006436444405052397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,12288,0.031833777825037636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,16384,0.2904862297905816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,10240,0.18715555138058135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,5120,0.09118044376373291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,16384,0.19832532935672334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,3584,0.150346663263109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,32,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,1024,0.008557333714432186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,256,0.055847108364105225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,4096,0.35261421733432347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,1024,0.007041778001520369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,1536,0.00692622239391009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,1536,0.007207999626795451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,3072,0.010785777535703448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,128,0.016304888659053378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,2048,0.015046222342385186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,1536,0.004679999831649992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,256,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,3584,0.12469777796003555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,512,0.0034133332471052804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,3584,0.18419910801781547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,2560,0.01885688967174954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,16384,0.1347608831193712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,768,0.024117334021462336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,65536,0.29643199178907603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,1536,0.05798844496409098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,768,0.007162666983074612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,16384,0.14229688379499647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,3072,0.06816177897983126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,1536,0.014039999908871122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,64,0.005066666752099991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,1536,0.004972444640265571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,51200,0.2738151020473904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,3584,2.965410656399197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,2560,0.044970668024486966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,3584,0.013772444592581855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,12288,0.32194754812452525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,51200,0.1639875570933024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,256,0.02203111184967889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,4096,0.06388355626000299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,16384,7.69830322265625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,7168,0.08440533611509536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,4096,0.34579022725423175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,2560,0.026314665873845417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,12288,0.45555024676852757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,16384,0.27422844039069283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,3072,0.11489955584208171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,512,0.0038080000215106537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,2048,0.01295644458797243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,128,0.007320000065697565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,16384,0.03185155656602647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,10240,0.1901235580444336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,3584,0.06812444660398695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,65536,0.6172169049580892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,7168,0.04840799834993151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,128,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,768,0.007102222078376346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,3072,0.08151733213000827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,3072,0.008841777841250101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,51200,0.27532000011867946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,1024,0.04997511042488945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,6144,0.0092613332801395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,51200,0.23982400364345977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,5120,0.04602844516436259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,16384,0.5362550947401259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,7168,0.008817777865462834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,8192,0.16452622413635254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,2560,0.012299555871221753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,12288,0.04756800002521939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,2048,0.5246328777737087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,12288,5.835115644666884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,32,0.007429333196745978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,5120,0.16350311703152126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,3584,0.011197333534558615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,12288,0.01996799972322252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,512,0.0037208890749348533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,12288,0.03810133205519782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,3584,0.007595555649863348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,64,0.0031191110610961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,4096,0.008757332960764566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,12288,0.17414489057328966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,7168,0.02728355593151516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,2560,0.007671111159854465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,51200,0.08426577515072292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,768,0.022498667240142822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,12288,0.018768888380792405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,256,0.005395555661784277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,1024,0.08723022540410359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,8192,6.872033013237847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,3072,0.0107360002067354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,3584,0.009743111001120673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,6144,0.033611555894215904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,768,0.028540445698632136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,128,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,1024,0.01146666705608368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,65536,0.40189067522684735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,768,0.0506568882200453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,5120,0.0129120002190272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,1024,0.004173333446184794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,12288,0.07978310849931505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,51200,0.07712088690863715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,768,0.005047111047638787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,6144,0.1220995585123698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,65536,0.1712986628214518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,65536,2.6685928768581815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,4096,0.03666844301753574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,5120,0.031194667021433514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,65536,0.3951893382602268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,3072,0.04011022382312351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,51200,0.12102044953240289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,7168,0.017044444878896076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,32,0.002667555585503578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,3072,0.06712710857391357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,8192,0.19105422496795654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,12288,0.13049599859449598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,10240,4.0678223503960504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,2048,0.0064693333374129395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,32,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,6144,0.013352889153692456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,65536,0.3681928846571181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,2560,0.02576711111598545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,2560,0.007961778177155389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,3072,0.01934400035275353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,128,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,128,0.0026737778551048706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,1024,0.43926133049858945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,64,0.03562577896647983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,6144,0.018566222654448617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,16384,1.1168444951375325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,64,0.06407288710276286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,256,0.005445333404673471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,8192,0.017278222574128043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,12288,0.29859556092156303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,128,0.003395555540919304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,32,0.012128888732857175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,65536,0.4893351131015354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,6144,0.1772906647788154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,65536,0.683234691619873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,8192,0.21481511327955458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,4096,0.0480275551478068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,7168,0.015170666906568738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,7168,0.027779555983013574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,5120,0.03141244583659702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,4096,0.11240177684360081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,7168,0.010774222513039907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,2048,0.02353511088424259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,7168,0.07589333587222628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,1536,0.0069431112044387394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,10240,0.3018426630232069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,2048,0.031445334355036415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,7168,0.018276444739765592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,128,0.008168000313970778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,16384,0.05049688948525322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,256,0.006520000182920032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,1536,0.02792000108295017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,2048,0.019359111785888672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,3584,0.00792266676823298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,1536,0.08563377459843953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,5120,0.04734399914741516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,4096,0.23240799374050566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,2560,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,1024,0.014085332552591959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,16384,0.11379555861155193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,256,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,12288,0.019761777586407132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,1536,0.024364444944593642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,3584,0.007400000260935889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,3584,0.3634924358791775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,3072,0.14143555694156223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,8192,0.02198666665289137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,8192,0.019863110449579027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,3072,0.008051555189821456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,256,0.03238044513596429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,512,0.007517333659860823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,512,0.0063573333124319715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,1024,0.00572533326016532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,12288,0.15758400493197972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,3072,0.018408889571825664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,256,0.005384888913896348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,6144,0.5549439854092068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,5120,0.150527106391059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,2560,0.017758222089873422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,1536,0.019306666321224637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,128,0.1861679951349894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,7168,0.051423999998304576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,5120,0.017873777283562552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,4096,0.015062222878138224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,6144,0.020947555700937908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,256,0.004963555683692296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,51200,0.2838097678290473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,512,0.031167109807332356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,10240,0.01440000037352244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,51200,0.09189866648779975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,64,0.0047733332547876565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,12288,0.11977510982089573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,256,0.003118222165438864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,4096,1.0119635264078777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,3072,0.020776889390415616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,512,0.006285333385070165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,32,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,4096,0.027423999375767175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,3584,0.051036443975236684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,6144,0.033710221449534096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,65536,0.31887910101148814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,3072,0.025656888882319134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,3072,0.006760889043410619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,4096,0.25057244300842285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,1536,0.004664888812435998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,2560,0.07317066854900785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,51200,1.392425749037001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,12288,0.13250222471025255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,4096,0.015895111693276297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,8192,0.4920506477355957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,8192,0.09437511364618938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,7168,0.014562666416168213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,12288,0.030386666456858318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,768,0.03674044542842441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,4096,0.25189420912000865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,4096,0.015712888704405892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,1024,0.021036444438828364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,51200,10.08992173936632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,5120,0.03197777933544583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,2560,0.013755555782053204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,512,0.019242667489581637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,1536,0.011864888999197217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,7168,0.013824889229403602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,8192,0.07368710968229505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,768,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,10240,0.07152889172236125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,4096,1.8438169691297743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,2560,0.027319111757808264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,51200,0.07564355267418756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,128,0.049435555934906006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,16384,0.02217866645918952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,512,0.007164444360468123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,128,0.00646577775478363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,10240,0.03152799937460158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,1536,0.01740622189309862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,32,0.007273777491516537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,32,0.012315555579132505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,2560,0.018972444865438674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,3072,0.06191111273235745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,12288,0.43293245633443195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,128,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,7168,0.02089244458410475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,2560,0.012689777546458773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,32,0.004396444393528832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,256,0.007748444047239091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,10240,0.18011732896169028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,512,0.00741155528359943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,2560,0.07006400161319308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,1536,0.01722400055991279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,12288,0.02825333343611823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,65536,0.7367839813232422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,65536,0.09512355592515732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,256,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,2048,0.012549333274364471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,16384,13.990723503960503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,3072,0.027408889598316614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,1536,0.08748800224728054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,4096,0.012837332983811697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,51200,4.041050804985894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,4096,0.1615768935945299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,1024,0.011746666497654386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,768,0.005922666854328579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,32,0.002971555623743269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,7168,0.18458488252427843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,64,0.005603555589914322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,16384,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,2560,0.005576888720194499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,256,0.0030666666312350165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,6144,0.0916408896446228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,2048,0.05938666396670871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,65536,1.437472873263889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,2048,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,51200,0.6217768987019857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,8192,0.36889065636528867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,32,0.01294844432009591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,51200,0.08503555589252049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,5120,0.010440888504187265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,16384,0.057357331116994224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,3584,0.07291644149356417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,12288,0.10156444708506267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,8192,0.03943644298447503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,51200,0.07878044578764173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,512,0.024197333388858374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,256,0.011065777805116443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,4096,0.06734222173690796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,768,0.18527732955084908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,16384,0.5820595423380533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,51200,1.6915067036946614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,8192,0.10275644726223415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,128,0.007263110743628607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,3072,0.09324444664849176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,3072,0.011770666473441653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,2048,0.011179555621412067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,7168,0.019114666514926486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,51200,0.9166204664442273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,3072,0.3411982324388292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,51200,0.3477804395887587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,51200,0.013049778011110095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,768,0.006791111081838608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,8192,0.09285155269834731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,2048,0.006649777707126405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,7168,0.12887733512454563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,3072,0.0919182234340244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,3072,0.04173600011401706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,12288,0.01719288859102461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,1536,0.014703111516104804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,512,0.0033573332346147964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,7168,0.10687110821406047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,10240,0.016686222619480558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,10240,0.024480889240900677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,12288,0.034747554196251765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,64,0.011988444460762871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,2048,0.0068195557428730865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,5120,0.020824889341990154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,32,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,65536,0.21793866157531738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,32,0.008571555217107138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,2560,0.18213244279225668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,2560,0.007095110913117726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,6144,0.02312622302108341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,768,0.0177848885456721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,6144,0.010002666877375709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,16384,0.9322417577107748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,6144,0.01886577738655938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,4096,0.06131466892030504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,2048,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,1536,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,2560,0.007239999870459239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,5120,0.023418666587935552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,8192,0.1089857750468784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,3072,0.10819911294513279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,3072,0.008975110948085785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,3584,0.026904889278941687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,65536,0.9589022530449761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,6144,0.050415999359554715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,2048,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,7168,0.007257777783605788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,65536,0.6350471178690592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,51200,0.11363111601935492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,3072,0.4456231329176161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,7168,0.031119111511442397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,512,0.017045333981513977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,1024,0.008112000094519721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,12288,0.44619199964735246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,3072,0.04160799913936191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,12288,0.1061297787560357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,128,0.004973333328962326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,65536,13.901920742458769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,65536,6.955457051595052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,1024,0.04410755634307861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,32,0.08394400278727214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,2048,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,768,0.005721777677536011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,12288,1.9757804870605469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,3584,0.011377777490350934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,256,0.005622222191757626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,32,0.006643555644485686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,51200,0.9578613705105252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,12288,0.5950346522861057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,512,0.0041582224269707995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,1536,0.007246221933099959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,51200,1.1406888961791992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,51200,0.4267608854505751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,3584,0.02740266587999132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,1536,0.006806222101052602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,512,0.014259555273585849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,65536,0.5442221959431967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,12288,0.07953066958321466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,1024,0.00499288903342353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,768,0.008011555506123437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,6144,0.5185653368631998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,5120,0.017488888568348356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,2048,0.0388995541466607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,8192,0.024863110648261175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,256,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,512,0.004965333475006951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,12288,0.03211288981967502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,10240,0.015375999940766228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,128,0.007735110819339752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,768,0.10303466849856907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,6144,0.2154559956656562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,7168,0.013648889131016202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,10240,0.029542220963372126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,2560,0.014087999860445658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,256,0.0031066667288541794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,12288,0.018615111708641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,65536,3.6281708611382375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,65536,0.29250044292873806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,1536,0.005903111149867375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,3072,0.17256800333658853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,3072,0.016756445169448853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,2560,0.03186044428083632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,51200,0.16398221916622585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,51200,0.020247111717859905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,3584,0.010620444185203977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,768,0.013164444102181328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,768,0.18408711751302084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,4096,0.024791111548741657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,256,0.030039999220106337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,256,0.003646222253640493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,5120,0.0358497765329149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,64,0.006956444846259222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,6144,0.008949333594905006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,128,0.004374222209056218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,2560,0.010652444428867765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,1536,0.019849777221679688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,65536,9.84108395046658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,16384,0.021056888831986323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,128,0.006615110983451207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,512,0.007618666523032718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,2560,0.013546666337384118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,16384,8.51114230685764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,256,0.0038293335172865125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,512,0.011666666302416058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,12288,0.5067111121283637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,128,0.03273422188229031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,512,0.0347013341055976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,10240,0.025943999489148457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,5120,0.17233778370751274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,5120,0.012487110992272695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,16384,0.344504886203342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,3584,0.09806844260957505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,32,0.015503999259736804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,1536,0.0047048889100551605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,7168,0.011257777611414591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,5120,0.010402667025725046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,8192,0.1372871134016249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,32,0.004072000169091755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,128,0.004334222111437055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,5120,0.02776799930466546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,12288,0.6453128920661079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,65536,0.18737688329484728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,3072,0.00869777798652649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,2560,0.008685333033402761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,4096,0.009692444569534725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,3072,0.009966221948464712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,64,0.12045155631171332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,2560,0.012823999755912356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,5120,0.05821777714623345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,768,0.07635288768344455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,10240,0.026341333985328674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,256,0.004218666503826777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,51200,2.3505261739095054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,4096,0.019684443871180218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,7168,0.008083555433485243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,256,0.011988444460762871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,16384,0.023038221730126276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,3072,0.011151110960377587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,12288,0.08116444614198473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,12288,0.02421155571937561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,16384,0.007915555602974361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,512,0.01828799976242913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,2048,0.6782053311665853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,64,0.004038222134113312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,10240,0.027111111415757075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,6144,0.02104533380932278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,1024,0.329678217569987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,2048,0.010876444478829702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,7168,0.03389244609408908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,4096,0.007298666569921706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,5120,0.013906665974193148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,7168,0.031888888941870794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,12288,0.3273013432820638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,1024,0.0057297779454125305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,7168,0.018716444571812946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,51200,0.2053013377719455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,256,0.012097777591811286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,8192,0.07623110877143012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,4096,0.02012977831893497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,1024,0.006878222028414409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,768,0.003584889074166616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,256,0.008153777983453538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,2048,0.007058666812049017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,32,0.003532444437344869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,2560,0.017095110482639737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,51200,0.2098888953526815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,768,0.003654222107595868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,3584,0.009311999711725447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,51200,0.040022220876481794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,16384,1.0760480033026802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,32,0.015523556205961438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,10240,0.1378177801767985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,512,0.07112355364693536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,256,0.005132444616821078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,128,0.17939378155602348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,10240,4.346554650200738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,12288,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,3072,0.022265776991844177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,3584,0.01128266668981976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,2048,0.015825778245925903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,7168,0.055289778444502086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,6144,0.005922666854328579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,2560,0.027680888772010803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,6144,0.006251555350091722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,51200,0.13893155256907144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,51200,0.33665956391228574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,256,0.0028595555987623003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,1536,0.011504000259770287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,768,0.01311288856797748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,10240,0.014473777678277759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,32,0.005664888769388199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,3072,0.31223376592000324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,5120,0.019461333751678467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,3584,0.008417777717113495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,32,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,7168,0.18600355254279244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,6144,0.010196444061067369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,768,0.025450666745503742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,768,0.017865777015686035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,1024,0.005506666584147348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,256,0.003116444374124209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,16384,0.038894222842322454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,8192,0.018392889036072623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,12288,0.018352889352374606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,128,0.009548444714811113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,128,0.004321777572234471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,5120,0.2807368967268202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,128,0.029764443635940552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,1536,0.011617778076065911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,2560,0.04864355590608385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,51200,0.05470755696296692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,8192,0.020204444726308186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,6144,0.014149333039919535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,16384,0.964023060268826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,10240,0.01956266661485036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,10240,0.028436445527606543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,64,0.004294222427739037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,128,0.0033608888172441055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,2560,0.014329777823554145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,7168,0.15913955370585123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,16384,6.532069312201606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,6144,0.02684266699684991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,8192,0.055063999361462064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,3584,0.4523164431254069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,1536,0.025423111187087163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,2560,0.006187555690606435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,7168,0.012668444050682915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,1024,0.004701333327425851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,12288,0.05393155415852865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,12288,0.08442044258117676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,256,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,3072,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,2560,0.13776532808939615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,16384,0.1028675569428338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,6144,0.021567111214001972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,4096,0.011249778171380361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,4096,0.01813333398765988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,256,0.015588444140222339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,256,0.0032746667663256326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,7168,0.0211173329088423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,128,0.018466666340827942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,2048,0.04876800047026741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,1024,0.004868444469239977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,2048,1.4179884592692058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,65536,0.26291023360358345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,768,0.025242666403452556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,3584,0.03309333324432373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,256,0.011687111523416309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,3072,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,4096,0.008615111311276754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,3072,0.016539555456903245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,16384,0.062430222829182945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,10240,0.0063768890168931745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,65536,0.690034654405382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,512,0.017685333887736004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,3584,0.12110043896569146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,1024,0.009142222503821055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,5120,0.2687404420640734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,7168,0.02048177851570977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,8192,0.028559111886554297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,1536,0.0070159998204973005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,2560,0.00869777798652649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,512,0.0058417779703934984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,32,0.002616000051299731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,3072,0.0073626670572492815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,5120,0.014478221535682678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,256,0.01128888875246048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,3072,0.0075288886825243635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,512,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,256,0.00517155561182234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,3584,0.03531466590033637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,32,0.008889777792824639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,2048,0.05858400132921007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,51200,0.16743466589185926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,128,0.00423555572827657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,65536,0.3772453202141656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,2560,0.009352888498041365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,32,0.027990221977233887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,10240,0.014021333720948962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,4096,0.009205333060688442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,1536,0.007319110963079665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,512,0.006164444403515921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,16384,0.02940800123744541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,5120,0.07683999670876397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,2560,0.007535999847782983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,51200,0.8130346404181586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,65536,2.171151902940538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,10240,0.01665511065059238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,10240,0.015152888165579902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,256,0.029008888536029395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,8192,0.015106666419241162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,16384,0.022882666852739122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,51200,24.16878594292535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,6144,0.010488000180986194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,128,0.0035004446076022256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,3072,0.013588444226317935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,65536,0.01609422266483307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,1536,0.006808888994985157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,128,0.005439110928111606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,768,0.013282666603724161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,16384,0.14635466204749212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,768,0.008629333641793992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,6144,0.021183111601405676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,8192,0.03785511189036899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,256,0.005480000128348668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,128,0.0035742223262786865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,64,0.008641777767075432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,7168,0.04749600092569987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,65536,0.40090489387512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,3072,0.007817777494589487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,1024,0.04919111066394382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,3072,0.022439110610220168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,8192,0.01627111103799608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,2560,0.09633244408501519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,2048,0.02151022189193302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,1024,0.011339555184046427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,8192,0.13277778360578749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,12288,0.046316444873809814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,5120,2.3300062815348306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,2560,0.044646223386128746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,256,0.013069333301650153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,5120,0.01699733402993944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,51200,0.6090862486097548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,8192,0.052689777480231396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,10240,0.04901955525080363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,7168,0.033287998702791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,51200,0.0636391109890408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,768,0.0036391110883818734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,5120,0.043545779254701406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,128,0.01223111069864697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,1024,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,3072,0.02791022260983785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,512,0.01313777764638265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,12288,0.0374311109383901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,5120,0.022996443841192458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,5120,0.011230222052998014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,128,0.005031110925806893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,64,0.006615110983451207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,512,0.022312889496485393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,256,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,3584,0.013542222479979197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,12288,1.416894170973036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,256,0.005791999813583162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,8192,1.1156667073567708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,3072,1.6715759701199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,10240,0.01069155583779017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,4096,0.37534578641255695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,256,0.03331466515858968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,12288,0.22332088152567545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,10240,0.184061328570048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,12288,0.015117333994971381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,5120,0.0248346659872267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,32,0.01593955523437924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,8192,0.08298933506011963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,12288,0.06100177764892578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,768,0.3753066592746311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,10240,0.03335466649797227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,16384,0.5064737531873915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,32,0.04155377878083123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,10240,1.6157564587063258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,512,0.006678222368160884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,12288,0.020322667227851022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,4096,0.06782311201095581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,65536,0.1303351057900323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,768,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,1536,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,6144,0.012221333053376941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,256,0.00703288863102595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,10240,0.13039555814531115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,4096,0.009390222529570261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,8192,0.015265777707099915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,51200,0.9061288833618164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,2560,0.0064195555945237475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,2048,0.013304000099500021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,5120,0.10437688562605117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,2048,0.008722667064931657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,1024,0.016957332690556843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,32,0.0028239999794297745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,2048,0.05331910981072319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,4096,0.12562578254275852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,1536,0.027951998843087092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,12288,0.6075119972229004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,16384,0.006280888699822956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,4096,0.060103999243842236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,7168,0.046457777420679726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,8192,23.054762098524307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,6144,0.11201777723100449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,16384,0.09264889028337266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,5120,0.027758222487237718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,4096,0.011643555429246692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,7168,0.3338026735517714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,5120,0.025366221865018208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,8192,0.1456071138381958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,10240,0.34212978680928546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,6144,0.9722426732381185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,3072,0.014857777290874057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,512,0.03494755427042643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,5120,0.013705777625242868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,7168,0.01577333278126187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,12288,0.6148631307813857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,64,0.0030826667530669104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,16384,0.0647182199690077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,2048,0.05299911234113905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,128,0.006986666884687211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,7168,0.04269599914550781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,5120,0.01888088881969452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,7168,4.304829491509332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,10240,0.011609777808189392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,512,0.019780443774329293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,128,0.0032319999817344877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,8192,0.1970942285325792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,3072,0.04187555445565117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,32,0.0036657779581016968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,3072,0.013660444153679742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,512,0.007911110917727152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,3072,0.006037333359320958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,10240,0.013625777430004544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,5120,0.10519999927944607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,8192,0.015588444140222339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,1024,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,512,0.00461422238085005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,7168,1.3932257758246527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,5120,0.13701866732703313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,2560,0.01296622223324246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,4096,0.00813066628244188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,6144,0.5502648883395725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,512,0.004239999999602635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,65536,0.7380782233344184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,12288,0.09403200281990899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,2048,0.010305778019958073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,256,0.016012445092201233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,2048,0.0075342224703894714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,12288,0.1671502192815145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,512,0.005783111270931032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,16384,0.43431999948289657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,768,0.01683200067943997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,2560,0.050287998384899564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,3072,0.0069004446268081665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,1024,0.02108000053299798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,64,0.01720266706413693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,1024,0.005413333161009683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,1536,0.04858044452137417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,65536,0.06076177623536852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,16384,0.2163404491212633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,65536,5.303093380398221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,12288,0.28691911697387695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,65536,0.18918400340610078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,4096,0.10519644286897446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,128,0.008468444148699442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,8192,0.03226844469706217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,51200,0.11105777819951375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,3584,0.02350666622320811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,6144,0.13243199719323054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,16384,0.34873421986897785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,5120,0.04862666792339749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,65536,0.0520702236228519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,5120,0.04733599887953865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,3072,0.008426666259765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,1024,0.013928888572586907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,128,0.24276089668273926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,1024,0.003995555556482739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,2560,0.007142222589916653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,3584,0.03369155526161194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,6144,0.016203555795881484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,1536,0.008336889247099558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,7168,0.014525334040323893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,5120,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,64,0.014039999908871122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,768,0.01351377781894472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,3584,0.011537777880827585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,12288,0.2632497681511773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,16384,0.1748951143688626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,256,0.009822222093741098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,16384,0.011117333339320289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,6144,0.006261333409282897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,10240,0.24372445212470162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,768,0.005712000032265981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,1024,0.008626666333940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,6144,0.008808000220192803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,512,0.02830222249031067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,512,0.0410844451851315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,2048,0.019778667224778067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,5120,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,65536,2.9888604482014975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,8192,0.11661067273881699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,512,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,10240,0.01581333412064446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,256,0.002955555501911375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,512,0.016141333513789706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,16384,0.0995146698421902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,128,0.004609777695602841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,16384,0.5200071334838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,768,0.012072000238630505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,51200,0.3532871140374078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,64,0.006214222146405115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,7168,0.23283465703328451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,1536,0.009684444301658207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,4096,0.01387466655837165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,128,0.0041031113101376426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,16384,0.019091554813914828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,2048,0.0517137779129876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,10240,0.08602222469117905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,65536,0.1793760061264038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,128,0.01551644504070282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,51200,0.08766400151782566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,128,0.011222222612963783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,3072,0.057767113049825035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,1536,0.005706666658322017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,16384,0.7049839761522082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,16384,0.05837511354022556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,8192,0.5275315708584255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,2560,0.23663287692599824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,2048,0.032607999112870954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,1536,0.05818577607472738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,7168,0.22982488738165963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,3584,0.0076142218377855085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,64,0.007974222302436829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,7168,0.014721777704026965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,768,0.005895111295912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,7168,0.01793511046303643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,2560,0.022783999641736347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,2048,0.019723556107944913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,2560,0.0069520001610120135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,16384,0.07934488852818807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,128,0.013095999757448832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,4096,0.02583911187118954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,7168,0.28687111536661786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,64,0.0058648888435628675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,4096,0.06271200047598945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,1536,0.006664888726340399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,1024,0.01904355486234029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,3072,0.025900445050663416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,256,0.0047128887640105355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,256,0.004861333303981357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,256,0.0033191111352708605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,5120,0.010057777994208867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,1024,0.010018666585286459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,3584,0.18364622857835558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,10240,0.016826666063732572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,6144,0.23818755149841309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,3072,0.011408000356621213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,32,0.0029582221888833572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,6144,0.011161777708265515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,6144,0.05759111377927992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,512,0.005821333163314396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,2048,0.019369777705934312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,10240,0.046132445335388184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,3072,0.023192000057962205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,12288,0.01918577816751268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,10240,0.006575999988449945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,51200,0.042223999897638954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,256,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,2560,0.021457778082953557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,32,0.0069475554757648045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,1024,0.20702844195895723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,1024,0.022515555222829182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,16384,0.04841599861780802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,51200,0.08269777562883165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,1024,0.010033778018421596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,512,0.00793599999613232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,768,0.05510488814777798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,768,0.003962666624122196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,512,0.005668444352017508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,5120,0.1339591079288059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,4096,0.10836533043119644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,768,0.02144266664981842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,16384,0.05203555689917671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,768,0.007495111061467065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,768,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,1024,0.012680889003806643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,2560,0.027572444743580286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,64,0.011791999969217511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,5120,0.006313777632183499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,3072,0.014863111906581454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,3584,0.01223111069864697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,64,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,12288,0.012837332983811697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,2560,0.031170666217803955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,3072,0.015132443772421943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,1024,0.05500088797675239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,1024,0.013473778135246701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,51200,1.361031108432346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,5120,0.02186222208870782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,10240,0.022896000080638464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,768,0.004815111143721474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,7168,0.04273510972658793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,4096,0.07187110847897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,32,0.0029724445194005966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,2048,0.05487111210823059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,16384,26.65808614095052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,51200,0.13273599412706164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,65536,0.12312889099121094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,512,0.0053839998112784485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,768,0.028387556473414104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,2048,0.018505778577592637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,3072,2.509138743082682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,1536,0.10988622241550022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,2048,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,4096,0.019158222609096102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,768,0.00500177757607566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,6144,0.008400888906584846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,1536,0.4051697783999973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,128,0.006996444529957241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,6144,0.01497688889503479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,16384,0.8037849002414279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,128,0.008654221892356873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,5120,0.0639431079228719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,65536,4.562126159667969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,10240,0.015226667126019796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,12288,0.009330666727489894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,64,0.03009866674741109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,1536,0.01127022256453832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,7168,0.019173334042231243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,64,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,3072,0.053611556688944496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,2560,0.031954665978749595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,64,0.005046222358942032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,64,0.004324444466167026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,7168,0.2318657769097222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,768,0.021384888225131567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,1536,0.7564204533894857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,65536,0.21583733293745253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,64,0.007473777565691207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,32,0.0028577778074476454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,1024,0.057463109493255615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,3072,0.007962666451931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,7168,0.35152533319261337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,1024,0.0058062221441004015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,256,0.005128889034191768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,12288,0.15194488896263972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,3072,0.010811555716726514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,2048,0.006545777950021956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,51200,0.0451164444287618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,7168,0.012298666768603854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,1536,0.01723911033736335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,1024,0.028949333561791316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,1536,0.013195555243227216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,3584,0.2768719991048177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,8192,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,3584,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,768,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,1536,0.00629688882165485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,32,0.005780444376998478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,256,0.0035662220584021676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,12288,0.02160177793767717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,10240,0.026563555002212524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,1536,0.006119110931952794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,6144,0.3358728885650635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,16384,0.25591643651326496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,5120,0.01831022236082289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,2560,0.007167999943097432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,1536,0.49189064237806535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,7168,1.1096933152940538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,512,0.03153422143724229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,256,0.00535733335547977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,2048,0.19314844078487822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,12288,4.4413981967502165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,1024,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,3584,0.010260444548394945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,5120,0.012804444465372296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,16384,0.0727377798822191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,7168,0.14376533031463623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,10240,0.08179555336634318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,6144,0.020097777247428894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,10240,0.04929866724544101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,2048,0.017444445027245414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,256,0.019448000523779128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,65536,0.08394133382373387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,8192,0.06381422281265259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,2560,0.05795822540918986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,8192,0.011743110915025076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,768,0.008760000268618265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,512,0.005593777944644292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,12288,0.9574399524264865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,4096,0.015824889143308003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,2048,0.051819556289248996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,2560,0.005978666659858491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,7168,0.09959110948774551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,7168,0.30176088545057506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,5120,2.0554871029324002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,256,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,32,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,65536,288.74197048611114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,3072,0.007826666865083907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,2048,0.030584888325797185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,1024,0.013658666776286231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,2560,0.06697511010699801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,2560,0.08143733607398139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,10240,0.7169822057088217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,51200,0.3811635441250271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,3584,0.011377777490350934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,10240,0.02760355505678389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,768,0.011335999601417117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,4096,0.009545777406957414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,51200,0.3373840120103624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,512,0.005967111223273807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,3584,0.00885688927438524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,10240,0.02537155482504103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,10240,0.22372976938883463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,3072,0.23716534508599174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,512,0.015610666738616096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,16384,0.03479289015134176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,16384,0.027219555444187585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,1024,0.03759288787841797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,3072,0.0355386667781406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,10240,0.045639998382992215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,64,0.0063884444534778595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,1024,0.009759999811649323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,65536,4.9730754428439665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,4096,2.257571538289388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,8192,0.051346666283077665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,8192,0.014460444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,256,0.0033760000434186724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,5120,0.016623111234770883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,7168,1.9705484178331163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,1024,0.02565066681967841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,51200,0.48011287053426105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,7168,0.10487466388278538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,2048,0.00980977796845966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,8192,0.02685866587691837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,64,0.00444177786509196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,8192,0.010136889086829292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,8192,0.08829244640138413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,128,0.005087999833954705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,10240,0.006413333531883027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,65536,0.7390809059143066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,5120,0.018048889107174344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,3072,0.02792888879776001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,256,0.003547555456558863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,6144,0.3191280100080702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,2048,0.09781333472993638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,5120,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,1536,0.1702737808227539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,4096,0.014148443937301636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,12288,0.03359288970629374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,5120,0.030254221624798242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,16384,0.04908711049291822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,128,0.007259555160999298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,3072,0.01107022249036365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,12288,0.015452444553375244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,4096,0.016248888439602323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,10240,0.0772071083386739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,3072,0.011350221931934357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,256,0.0047404443224271136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,5120,0.008935111264387766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,5120,0.23775466283162436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,5120,0.1690115531285604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,8192,0.012031111452314587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,10240,0.02291733357641432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,7168,0.011376000112957425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,7168,0.08815733591715495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,6144,0.016338666280110676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,768,0.010215111076831818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,1024,0.01314222233162986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,768,0.02211644417709774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,10240,0.031494223409228854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,3072,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,64,0.005327111317051782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,7168,0.011139555937714048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,768,0.011075555450386472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,64,0.0058471109304163195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,768,0.0036595554815398324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,768,0.08322311109966703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,64,0.006814222368929122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,2560,0.015808888607554965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,4096,0.0321760012043847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,51200,0.30459290080600315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,2560,0.015094222293959724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,64,0.0026346666531430352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,51200,0.34978400336371523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,12288,0.018499554859267343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,6144,0.010223999619483948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,65536,0.17464355627695718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,16384,0.01223288890388277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,768,0.048678222629759044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,5120,0.046295109722349376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,2048,0.01332800007528729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,12288,0.16518310705820718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,8192,0.14957333935631648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,3072,0.016508445143699646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,128,0.00282222218811512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,3072,0.01074577785200543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,512,0.2599484390682644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,32,0.02179644505182902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,2048,0.010673777924643623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,65536,0.9915164311726888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,2560,0.9481804105970594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,8192,0.02230311102337307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,128,0.007695111135641734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,2560,0.049379554059770375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,1536,0.008348444269763099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,1024,0.014852444330851236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,5120,0.012424888710180918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,32,0.006554666492674086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,7168,0.038896001047558255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,7168,0.011185777684052786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,5120,0.013263999587959714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,3584,0.007045333584149678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,8192,0.015415999624464246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,10240,0.016632889707883198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,32,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,2560,0.006213333457708359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,8192,0.011768889096048145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,3584,0.008821333448092142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,12288,0.8600311279296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,51200,1.1167484919230144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,16384,4.025742424858941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,2048,0.06621955500708686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,256,0.005929777605666055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,768,0.006367110957702001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,768,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,3584,0.09415911303626166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,256,0.00378577783703804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,1536,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,1024,0.0064088888466358185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,4096,0.5109368960062662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,32,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,65536,0.2960275544060601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,51200,13.428988138834635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,4096,0.038571556409200035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,51200,0.16854577594333223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,2560,0.03235022226969401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,51200,0.23202400737338594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,8192,0.23799644576178658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,1536,0.009549332989586724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,128,0.004330666528807746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,768,0.029926220575968426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,5120,0.2873377799987793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,12288,0.024660444921917383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,64,0.0026195556339290407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,64,0.002567111204067866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,32,0.005374222166008419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,4096,0.013216888739003075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,4096,0.01993688941001892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,256,0.019340443942281935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,3584,0.19820177555084229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,65536,0.16775288846757677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,1024,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,128,0.005286222116814719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,768,0.0034675554682811103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,6144,0.012323555847009024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,12288,0.03819377885924445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,65536,0.10162666771146987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,16384,0.23268445332845053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,64,0.002599999929467837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,256,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,2560,0.012347555822796293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,3584,0.10229244497087266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,3072,0.009488888912730748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,65536,9.967323303222656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,2048,0.007167999943097432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,7168,0.2898622353871663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,768,0.005848000033034219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,7168,0.018532445033391316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,128,0.011505777637163797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,64,0.00536088893810908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,1024,0.04141955574353536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,1536,0.04255733225080702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,5120,0.017403556240929496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,5120,0.032163557079103254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,256,0.003080000066094928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,64,0.0053013331360287136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,65536,4.104966905381945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,1536,0.00778311077091429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,512,0.008418666819731394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,12288,0.04012799925274319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,10240,0.9172852834065756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,65536,1.2538426717122395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,51200,3.718242645263672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,768,0.012650666965378655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,16384,0.47436979081895614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,8192,0.00849422232972251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,3584,0.3536835511525472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,1536,0.010320888625250923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,16384,0.2398782306247287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,1024,0.016435555285877652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,16384,0.025942222939597234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,3584,0.0615288880136278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,5120,7.001380496554905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,3584,0.18572978178660074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,512,0.004442666553788715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,3584,0.006817777951558431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,3584,0.008709333009190029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,7168,0.010381333529949188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,64,0.0028373334142896864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,16384,0.05730666716893514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,10240,2.767048094007704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,7168,0.14479555024041071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,512,0.004062222109900581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,12288,0.019147555033365887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,256,0.029678222205903795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,2048,0.09145155880186294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,64,0.01683644453684489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,768,0.012588444683286877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,2048,0.009753777749008602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,4096,0.030024889442655776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,10240,0.06021777788798014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,16384,0.019690665933820937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,128,0.0070364442136552595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,65536,0.040559109714296125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,2048,0.006794666664467917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,32,0.02253866692384084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,64,0.0073831114504072405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,768,0.005658666706747479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,65536,0.20381244023640951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,2048,0.011741333537631564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,51200,2.2534205118815103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,256,0.01347555551264021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,65536,0.14047822687360975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,51200,0.41470932960510254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,16384,0.05005333489841885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,12288,0.045138667027155556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,128,0.022133333815468684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,6144,0.01204622205760744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,5120,0.15158132712046304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,51200,0.3769991132948134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,7168,0.014536889062987434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,128,0.004876444323195351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,8192,0.06432622008853488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,32,0.005095110999213325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,512,0.01033155537313885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,6144,0.01720355616675483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,1024,0.054902222421434194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,65536,0.08057955238554212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,2560,1.52655824025472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,128,0.003768889026509391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,6144,0.01637066735161675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,3584,0.2046675549613105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,5120,0.0826337801085578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,4096,0.030307554536395605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,256,0.16474666860368517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,6144,0.08453955252965291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,65536,3.8073175218370228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,1536,0.030806223551432293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,1536,0.012671999633312225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,2048,0.423050668504503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,3072,0.03333777851528592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,6144,0.15565778149498835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,512,0.08682933118608262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,6144,0.02440622283352746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,4096,0.05411466625001696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,1536,0.00609422226746877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,128,0.004411555412742826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,32,0.002584888910253843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,3584,0.020579554968410067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,7168,0.1700293355517917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,51200,0.48541869057549375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,2560,0.13022755251990423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,1536,0.2727271185980903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,256,0.008755555583371056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,3584,0.15859466128879124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,5120,0.22597156630622017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,128,0.002959111084540685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,4096,0.18790133794148764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,3584,0.04596977763705783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,1024,0.03143289022975498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,32,0.005483555710977978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,1536,0.004855999930037392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,64,0.0029582221888833572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,6144,0.06439022223154704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,7168,0.027103111147880554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,7168,1.087908426920573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,16384,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,1024,0.29093954298231334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,32,0.006681777950790193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,3584,0.01758311192194621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,5120,0.07647822300593059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,5120,0.016249777542220224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,256,0.007894222107198503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,2048,0.0068764446510208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,6144,0.006383111079533895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,10240,0.07942844761742486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,10240,1.2973591486612956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,16384,0.21465067068735758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,7168,0.3316711054907905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,8192,0.20661777920193145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,256,0.004931555440028508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,12288,0.01238044434123569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,10240,0.036665777365366616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,65536,0.2980417675442166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,16384,0.019391111201710172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,65536,0.0636319981680976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,7168,0.03475022315979004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,8192,0.01257333325015174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,10240,3.7715047200520835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,2048,0.07521510786480375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,65536,0.42945334646436906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,10240,0.15967110792795816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,16384,0.08981777562035455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,12288,0.015786666009161208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,12288,0.03861600160598755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,32,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,128,0.004034666551484002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,2560,0.007096889118353526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,2048,0.0069671107663048645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,2048,0.006789333290523953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,256,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,51200,3.218783908420139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,256,0.006996444529957241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,51200,0.18723289171854654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,2560,0.13074399365319148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,3072,0.011944889194435544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,16384,0.16550666756100124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,5120,0.03184355629814996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,1024,0.004418666578001446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,256,0.011035555766688453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,51200,3.992510053846571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,768,0.01312355531586541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,5120,0.008849778109126622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,3584,0.008343111309740279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,10240,0.017264889346228707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,6144,0.010654222634103564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,6144,0.019172444939613342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,10240,0.02803200152185228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,2048,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,2560,0.022340445054901972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,10240,0.0768231087260776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,512,0.005201777650250329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,512,0.007975111405054728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,7168,0.13865955670674643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,1024,0.0064488889442549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,2048,0.006806222101052602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,65536,33.55415513780382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,128,0.10519111156463623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,32,0.002814222127199173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,2048,0.21051377720303008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,12288,0.203165332476298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,2560,0.047107554144329496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,3072,0.008386666576067606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,8192,0.021896888812383015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,512,0.01312355531586541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,2560,0.0688213308652242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,5120,0.26185690032111275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,6144,0.07133066654205322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,768,0.013252444565296173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,4096,0.07878577709197998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,64,0.005642666584915585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,16384,0.07351733578575982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,51200,0.07055911090638903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,65536,2.3240345848931208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,10240,0.1373751163482666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,10240,0.0822311109966702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,768,0.02304711110062069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,3072,0.2448666625552707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,512,0.008015111088752747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,51200,0.230858670340644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,65536,0.52347289191352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,65536,0.09174844291475083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,256,0.026498667067951624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,12288,0.03394755721092224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,16384,0.044362667534086436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,2560,0.007850666840871176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,5120,0.05359733435842726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,1024,0.03199022346072727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,1536,0.01255022237698237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,32,0.019222221440739103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,2560,0.022329777479171753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,1024,0.07852977514266968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,8192,0.02375999920898014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,8192,0.009129777550697327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,6144,0.5546115769280328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,256,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,65536,0.10871644152535333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,4096,0.010634666515721215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,768,0.005758222192525864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,7168,0.02544266647762722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,1536,0.015896888242827523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,5120,0.031743112537595965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,7168,0.4817164209153917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,8192,0.02215733296341366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,65536,2.861385769314236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,16384,1.2309306462605794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,3584,0.031509333186679415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,1536,0.2094951073328654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,51200,0.6699413193596734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,51200,8.944441901312933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,7168,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,10240,0.28520798683166504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,512,0.00518133325709237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,5120,0.015027556154463025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,32,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,768,0.010588444769382477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,3072,0.019186667270130582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,16384,1.1485040452745225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,2048,0.429275565677219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,1536,0.0059164443777667145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,65536,0.16977155208587646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,128,0.004537777768241035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,65536,0.13808622625139025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,7168,0.03895644346872965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,64,0.006664888726340399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,16384,0.07947911156548394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,256,0.003056888985964987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,2560,0.011330666641394297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,51200,2.155331505669488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,2048,0.07119199964735243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,65536,0.13672888278961182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,4096,0.06212888823615181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,4096,0.006126222097211414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,128,0.004367111043797599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,2048,0.014520888527234396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,16384,0.0165955556763543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,128,0.005510222166776657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,10240,0.013589333328935834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,8192,0.10915288660261367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,10240,0.23491999838087294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,7168,0.031122667921913996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,128,0.01704355577627818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,64,0.0038000001675552795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,6144,0.4253982173071967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,8192,0.11759911643134223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,8192,0.10358222325642903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,8192,0.1005786657333374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,7168,0.023603555228975084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,51200,0.2809573279486762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,7168,0.012870222330093384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,2560,0.9905253516303169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,7168,0.0166631109184689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,512,0.004449777719047335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,1536,0.007648888561460707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,4096,0.6320711241828071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,4096,0.014857777290874057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,512,0.022387555903858606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,6144,0.012495111260149213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,6144,0.025250666671329077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,7168,0.010862222148312463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,6144,0.030637333790461223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,32,0.006199999815887875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,768,0.005032888717121548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,10240,0.5835422409905328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,3584,0.04408533374468485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,12288,0.03107288810941908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,256,0.013925333817799887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,7168,0.007183111376232571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,1024,0.024559999505678814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,1536,0.010676444404655032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,3584,0.007981333467695449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,2560,0.008030222521887885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,12288,0.7195679876539441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,1536,0.030913776821560327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,2048,0.005376889059940974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,1024,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,2048,0.10566577646467422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,10240,0.10486400127410889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,32,0.003335111050142182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,256,0.0029066666546795103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,8192,0.019144889381196763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,32,0.0032675556010670135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,8192,0.03983733389112685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,5120,0.032143112685945295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,2560,0.01015377789735794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,8192,0.015057777365048727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,5120,2.5523192087809243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,2560,0.012664000193277994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,51200,0.2096924384435018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,128,0.00868800034125646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,256,0.007499555746714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,2560,0.038981331719292536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,6144,0.013813333378897773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,3072,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,3584,0.0068862222962909276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,5120,0.2146915594736735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,256,0.007438222567240397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,512,0.009565333525339762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,12288,0.9303724500868055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,2048,0.021161778105629817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,4096,0.01312622262371911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,1024,0.38989689615037704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,5120,0.03745066788461473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,7168,0.005986666513813867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,32,0.0026382222357723448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,1536,0.006276444428496891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,512,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,12288,0.006803555621041193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,512,0.01647911138004727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,65536,1.304676479763455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,2560,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,1024,0.005752000129885144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,6144,0.39097955491807723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,10240,1.9071449703640406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,65536,0.1796791156133016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,3584,0.03987733191914029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,10240,0.05271466573079427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,768,0.007870222131411234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,512,0.00583022211988767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,768,0.2186328834957547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,2048,0.01885422236389584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,1024,0.02571288910177019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,2560,0.006021333237489064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,51200,9.158643934461805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,4096,0.1139413383271959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,10240,4.716388278537326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,5120,0.020473778247833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,32,0.0041271112859249115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,5120,0.015625778171751235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,1024,0.005976888868543837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,2048,0.2985546588897705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,12288,0.027136888768937852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,128,0.016466667254765827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,256,0.02403555479314592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,6144,0.40165334277682835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,3584,0.009844444692134857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,16384,3.855175018310547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,2048,0.00906666699382994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,512,0.034880889786614314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,128,0.013016000390052795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,3584,0.12465066379970974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,5120,0.012170666621790992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,1024,0.003955555458863576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,12288,0.07946577999326918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,512,0.005504888792832692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,3584,0.017242666747834947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,3072,0.07291822301016913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,6144,0.05422755413585239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,12288,0.009957333405812582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,768,0.011437333292431302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,4096,0.0326222214433882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,32,0.00240000006225374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,256,0.1586151123046875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,7168,0.0853653351465861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,5120,0.14184266991085476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,51200,0.06287733051511976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,128,0.034416887495252825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,1536,0.01625422139962514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,5120,0.023426666855812073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,7168,0.07978222105238172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,16384,0.027102222045262653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,2560,0.0099973330895106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,2560,0.04996177885267469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,3072,0.15808799531724718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,7168,0.11836532751719157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,4096,0.00620266670982043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,256,0.003825777934657203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,2560,0.018789332773950364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,256,0.0068808889223469635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,1024,0.006643555644485686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,3072,0.2824595504336887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,32,0.033963554435306124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,3584,0.06456266509162055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,2560,0.007249777515729268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,8192,0.017422222428851657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,64,0.007131555842028723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,8192,0.0211182220114602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,51200,1.2780773374769423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,3584,0.027756444282001917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,5120,0.008791111409664154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,256,0.00794222205877304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,1536,0.006346666564544042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,768,0.03176800078815884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,256,0.005568888866239124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,128,0.02330488959948222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,6144,0.131113780869378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,12288,0.22430133819580078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,2048,0.005058666484223472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,2048,0.027829332484139338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,51200,1.4041190677218969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,1536,0.07350133525000678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,7168,0.01220355596807268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,3584,0.008647110727098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,1024,0.01089511149459415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,16384,0.026720888084835474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,1536,0.2124320003721449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,3072,0.011476444701353708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,10240,0.18917244010501436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,256,0.23162134488423666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,7168,0.1371848848130968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,2560,0.12789066632588705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,2560,0.008061333662933773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,1024,0.03501510951254103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,51200,4.494510226779513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,256,0.004231111043029361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,7168,0.01590844492117564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,4096,0.027422222826215956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,5120,0.01776000029510922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,6144,0.06241866615083483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,512,0.019152889649073284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,2560,0.01573600040541755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,16384,0.5643262333340114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,2048,0.10888977845509847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,128,0.004552000098758274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,64,0.0031715554909573663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,256,0.0057297779454125305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,16384,0.445983091990153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,768,0.010290666586822933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,5120,0.04412800073623657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,8192,0.1551768912209405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,4096,0.009703999592198266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,16384,0.7436933517456055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,16384,0.2731920083363851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,64,0.002943111169669363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,1024,0.012459555433856117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,768,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,10240,0.49988889694213867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,128,0.008816888762844933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,1024,0.01035911093155543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,7168,0.01601777805222405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,5120,0.009721777505344814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,5120,0.023232888844278123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,1024,0.012311999996503195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,7168,0.033024887243906655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,256,0.026421333352724712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,8192,0.0148062225845125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,10240,0.0776124464140998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,1024,0.01759377784199185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,51200,0.7572773297627767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,16384,2.630354563395182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,128,0.004960000101062986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,1536,0.047091556919945605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,1536,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,12288,0.3285733328925239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,1536,0.015032889114485847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,7168,0.037837333149380155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,3584,0.020578665865792167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,5120,0.1692239973280165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,3584,0.09414133098390366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,1536,0.011672000090281168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,10240,0.013106666505336761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,12288,0.024687111377716064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,12288,0.061181333329942494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,3584,0.07181777556737264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,8192,0.156131558948093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,12288,0.018792889184421964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,2560,0.1781217787000868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,64,0.002624888800912433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,65536,0.0686640010939704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,8192,0.05092622174157036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,256,0.007304000357786815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,5120,0.00960355583164427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,256,0.0032666667054096856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,768,0.0034391110142072043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,4096,0.017765333255132038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,8192,3.3652960459391275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,4096,0.020455999506844413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,64,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,32,0.02602844436963399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,32,0.0027991111079851785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,6144,0.35828977160983616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,256,0.005690666536490123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,256,0.006865777903132969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,64,0.020225778222084045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,4096,0.014707555373509726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,16384,0.05553599860933092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,6144,0.02628266645802392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,1536,0.0129439996348487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,5120,0.015607110328144498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,128,0.00906399968597624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,8192,0.008984000318580205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,32,0.020978666014141504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,7168,0.04275466667281257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,3584,0.007381333245171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,4096,0.033039109574423894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,2560,0.018615111708641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,3584,0.04057866666052077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,8192,0.8083173433939616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,7168,0.021838222940762837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,10240,0.2248577806684706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,512,0.017462222112549674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,7168,0.3759875562455919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,4096,0.029913776450686987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,768,0.03951199849446615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,12288,1.105447981092665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,1536,0.008162666526105668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,10240,0.0702986650996738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,4096,0.015420445137553744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,65536,0.4920933511522081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,5120,0.5074231359693739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,5120,2.5575154622395835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,16384,0.052667554881837636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,128,0.018187556001875136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,12288,0.038557334078682795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,7168,0.020983111527231004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,256,0.015592000550693937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,16384,0.5249626901414659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,128,0.0027626665929953256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,32,0.008761777646011775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,1024,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,128,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,16384,0.02435644467671712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,1024,0.06395822101169162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,768,0.0038897775941424896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,12288,0.02697244452105628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,32,0.0029680000411139596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,8192,1.7842169867621527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,2048,0.008053333395057255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,3072,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,2048,0.011940444509188334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,4096,0.20288356145222983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,1024,0.012346666720178394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,51200,3.7040969000922312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,1024,0.01015377789735794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,5120,0.2816889021131727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,2048,0.02109688851568434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,128,0.009291555318567488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,12288,0.019154666198624503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,1536,0.02752977775202857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,2048,1.3127146826850045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,512,0.0041235557032956015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,128,0.005081777771313985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,768,0.005538666827811136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,7168,0.015054222610261707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,3584,0.032215111785464816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,2048,0.012256888879670037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,4096,0.04348266786999173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,6144,0.010258666343159145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,32,0.002603555512097147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,6144,0.0821288890308804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,51200,0.05416355530420939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,128,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,4096,0.3041786617702908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,7168,0.016758221719000075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,65536,0.2685840129852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,256,0.1324453353881836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,6144,0.05934400028652615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,65536,0.13832444614834255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,16384,0.19730488459269205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,3072,0.007469333708286285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,32,0.1922995514339871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,8192,0.013927110367351107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,65536,1.742711067199707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,128,0.0027262222849660446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,1024,0.008655110994974772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,256,0.03794222076733907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,2560,0.017883555756674874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,2048,0.014286221729384528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,4096,0.02569155560599433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,512,0.005685333162546158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,512,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,1024,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,3072,0.013579555683665805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,6144,0.04378844300905863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,12288,0.018848000301255118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,512,0.009305777649084726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,32,0.004597333156400257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,2560,0.061054223113589816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,3584,0.013017777767446307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,1024,0.05310044354862637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,4096,0.03429777754677667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,7168,18.689090304904514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,4096,0.07913511329227023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,12288,0.3042515648735894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,3584,0.08601511187023586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,51200,0.06472444534301758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,4096,0.010996444357766045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,16384,0.021707556313938562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,16384,0.33554577827453613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,4096,0.006493333313200209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,768,0.007807999849319458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,3584,0.006846222198671765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,2048,0.012033777932325998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,12288,0.015947555502255756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,512,0.003699555579158995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,51200,1.251391092936198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,2048,0.012810666528013019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,1536,0.008902221918106079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,12288,6.7266053093804254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,768,0.0047395556337303585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,65536,0.1158871120876736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,10240,0.11560266547732884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,3584,0.05273511012395223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,64,0.026349332597520616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,1024,0.016398222910033334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,5120,0.43008889092339414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,5120,0.012995555169052549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,1536,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,3072,0.011576000187132092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,65536,0.36511021190219456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,3584,0.18292533026801217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,16384,0.022080888350804646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,12288,0.030557334423065186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,1536,0.06886311372121175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,768,0.10715111096700032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,6144,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,7168,0.14159377415974936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,32,0.12824532720777723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,51200,0.9853191375732422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,2560,0.06112088759740194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,2560,0.014728888869285583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,64,0.006599111275540457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,3584,0.008023999631404877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,3072,0.024158222807778254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,512,0.1195333268907335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,6144,0.023732443650563557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,6144,0.026387555731667414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,12288,0.19242666827307808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,2560,0.010112889111042023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,6144,0.12259644932217068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,64,0.002560888934466574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,7168,0.08099289072884454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,7168,0.03417511118782891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,256,0.003633777714437909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,65536,0.03136000037193298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,768,0.010824888944625854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,128,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,5120,0.01720355616675483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,12288,0.6162257724338108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,6144,0.17227555645836723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,64,0.002605333303411802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,16384,0.052073776721954346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,768,0.005632888939645555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,10240,0.8341937594943576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,4096,0.11166400379604763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,51200,13.022400750054253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,7168,0.03340977761480544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,64,0.004776000148720211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,5120,0.022494221727053326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,16384,0.07215911149978638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,5120,0.049682666858037315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,51200,0.22477955288357207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,2560,1.340879970126682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,51200,0.5223271052042643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,128,0.004019555532270008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,16384,0.04792266752984789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,8192,0.26114487648010254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,12288,0.017663111289342243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,2048,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,4096,0.02295822236273024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,8192,0.011639110743999481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,65536,2.38484615749783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,128,0.02091377807988061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,3072,0.011759999725553723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,128,0.012151111331250934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,65536,0.8842373424106174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,256,0.08767910798390706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,65536,0.10308444499969482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,12288,0.24705510669284395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,3584,0.10247555706236099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,768,0.008943999807039896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,3584,0.009201777478059133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,7168,0.03194933467441135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,512,0.0034053333931499054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,10240,0.006552000012662675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,4096,0.055445333321889244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,12288,0.055797331862979464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,1024,0.2902711232503255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,51200,4.314646402994792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,2048,0.009373333719041612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,10240,0.08086578051249187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,128,0.010625777973069085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,512,0.017895110779338412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,7168,0.18488800525665283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,3584,0.015559110376569959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,6144,0.013702222042613559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,768,0.006766222417354584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,1024,0.012302222351233164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,128,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,6144,0.27957955996195477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,51200,0.23291556040445963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,6144,0.33946577707926434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,3584,0.022095999783939783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,16384,0.542102230919732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,768,0.009153777526484596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,2048,0.09621333413653904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,768,0.006240889016124938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,128,0.016389333539538913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,16384,0.09109244743982951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,12288,0.3411448796590169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,16384,0.05169777737723457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,51200,0.9505333370632596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,6144,0.012898666991127862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,10240,0.01794755624400245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,3072,0.18927021821339926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,512,0.006077333456940121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,1536,0.08180177874035306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,8192,0.03127288818359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,64,0.005997333261701796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,2048,0.019392000304328073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,65536,0.5391679869757758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,2048,0.05652888615926107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,128,0.009684444301658207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,6144,0.05377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,4096,0.024259555670950148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,3584,0.2481022146013048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,12288,0.015115555789735583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,12288,0.15241866641574436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,65536,0.7390595542060004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,5120,17.18712700737847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,51200,0.07759288946787517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,10240,0.04065777858098348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,1024,0.048031111558278404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,8192,0.01607288916905721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,4096,0.07779022057851155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,7168,0.09647822380065918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,2560,0.009480000370078618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,32,0.005252444495757421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,64,0.0028444443725877335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,2048,0.009269333548016017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,6144,0.3809466626909044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,6144,0.010437332921557956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,3584,0.11829866303337945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,10240,0.01774044500456916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,2560,0.07182577583524916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,3072,0.005895111295912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,512,0.018962666392326355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,768,0.024363555841975745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,1536,0.009140444298585257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,2560,0.013325333595275879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,256,0.0053013331360287136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,2560,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,16384,0.04907822277810839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,4096,0.06766666968663533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,8192,0.012614222036467658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,1024,0.025015999873479206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,128,0.003947555604908201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,256,0.0328693323665195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,8192,0.13367999924553767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,7168,0.01924444403913286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,1536,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,1536,0.010063110954231685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,7168,0.021720889541837905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,2048,0.008838222258620793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,5120,0.027681777874628704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,512,0.030511998467975195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,1536,0.006370666540331311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,3072,0.012011555333932241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,3072,0.05746044715245565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,3072,0.05890311135186089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,1536,0.020090666082170274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,3072,0.009944000177913243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,7168,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,8192,0.05718666977352566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,8192,0.021055110626750525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,1536,0.011477332976129321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,10240,0.015551111764378019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,1024,0.1999280055363973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,3072,0.017685333887736004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,32,0.006865777903132969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,2048,0.008599999878141616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,2048,0.008577778107590145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,64,0.01053955571519004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,6144,0.04675733380847507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,128,0.011529777612951068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,2560,0.014871110518773397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,32,0.009458666874302758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,2048,0.018990221950742934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,6144,1.5166346232096355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,256,0.003572444534964032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,2560,0.010378667049937779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,768,0.0058222222659322954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,1024,0.006492444624503453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,6144,0.3091742197672526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,7168,2.5985590616861978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,256,0.02104266650146908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,32,0.007408888803588019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,32,0.009455111291673448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,12288,0.02621777852376302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,768,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,5120,0.023020444644822016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,8192,0.011529777612951068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,3584,0.06705155637529162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,3072,0.018671999375025432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,12288,0.26495112313164604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,512,0.003742222156789568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,512,0.015786666009161208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,2560,0.09030755360921223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,7168,0.05817777580685086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,512,0.0035093331502543557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,256,0.02258133391539256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,768,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,128,0.006711111300521427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,16384,0.021212443709373474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,3072,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,6144,1.5513582229614258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,4096,0.011201777391963534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,1536,0.03336888882848952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,8192,0.01404622197151184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,12288,0.02158044444190131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,3584,0.026686223016844854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,6144,0.0178195552693473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,2560,0.01735111077626546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,64,0.021295999487241108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,5120,0.009973333113723332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,1536,0.0177839994430542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,12288,0.057525330119662814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,64,0.0033084443873829315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,2048,0.006302222195598815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,2560,0.013491555220550962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,12288,0.012900444368521372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,768,0.014279999666743808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,1536,0.011233777635627322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,64,0.0069546666410234236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,8192,0.013872000078360239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,768,0.004271999829345279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,2560,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,5120,0.010650667051474253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,2560,0.010168000227875179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,1536,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,2560,0.012162666353914471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,12288,0.03888266616397434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,65536,1.954482608371311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,8192,0.07223733266194661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,7168,0.02161955502298143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,6144,0.06263911061816745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,256,0.012511110968059964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,64,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,32,0.007053333024183909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,2560,0.10078222221798366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,6144,0.00775733341773351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,7168,0.010774222513039907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,5120,0.1830995612674289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,7168,0.1965555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,3584,0.006703111032644908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,768,0.007874666816658443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,5120,0.011911110745535957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,3072,0.015208888385030957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,768,0.0038311113086011675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,512,0.004930666751331753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,12288,0.2607653405931261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,256,0.04283555679851108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,4096,0.01479022204875946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,128,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,128,0.0046533333758513136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,16384,0.7610479990641276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,16384,0.16659555170271131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,3072,0.01677688956260681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,51200,0.19497244887881807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,16384,0.06913866599400838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,128,0.028130667077170476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,512,0.0034835553831524322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,3072,0.00868888861603207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,768,0.04128177629576789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,256,0.03515999846988254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,3072,0.00793511089351442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,5120,0.022278222772810195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,512,0.005033777819739448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,7168,0.014122666584120857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,8192,0.01200355589389801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,7168,0.014035556051466199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,3072,0.008348444269763099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,7168,0.01532977819442749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,16384,0.02257688840230306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,7168,0.18751555018954805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,1536,0.03559911251068115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,128,0.002963555562827322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,7168,0.029706666866938274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,768,0.0035662220584021676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,128,0.002741333304180039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,128,0.004518222063779831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,512,0.00332266671790017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,16384,0.021337777376174927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,128,0.007648000286685095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,1536,0.004393777913517422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,16384,0.21179555522070992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,10240,0.2366711033715142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,2560,0.007747555772463481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,768,0.003843555433882607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,4096,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,3072,0.5409582455952963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,10240,0.07537955707973905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,512,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,12288,0.027706666125191584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,65536,0.9901492860582141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,2048,0.09977866543663873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,16384,0.10779377486970688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,512,0.04229244589805603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,5120,0.02048000031047397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,10240,0.2471680111355252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,1024,0.007267555428875818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,10240,0.06663378079732259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,7168,0.014541332920392355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,3584,0.007232888705200619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,3072,0.008170666793982187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,4096,0.03641866644223531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,4096,0.011388444238238864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,12288,10.447835286458334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,5120,0.014894222219785055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,128,0.5840053558349609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,32,0.004501333253251182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,16384,0.027173333697848853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,6144,0.15994221634334987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,1536,0.007012444237867991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,4096,0.011583111352390714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,3072,0.022444445225927565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,1024,0.14229599634806314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,7168,0.01455200049612257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,16384,0.030603554513719346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,512,0.020473778247833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,3584,0.03256533212131924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,51200,0.8730373382568359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,1536,0.055959999561309814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,3584,0.01037688884470198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,10240,0.6172729068332249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,3584,0.017231111725171406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,5120,0.033655110332700945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,1024,0.020255999432669747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,768,0.00702044450574451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,128,0.008912888665994009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,4096,0.01203555530971951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,256,0.00349866681628757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,128,0.003125333330697484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,1536,0.005201777650250329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,128,0.005192888693677055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,2560,0.08533422152201335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,512,0.04977777931425306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,2048,0.01735111077626546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,5120,0.013458666702111563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,2560,0.011751111182901593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,1024,0.007232888705200619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,256,0.007292444507280986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,3584,0.009304000271691216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,16384,0.025120000044504803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,64,0.004014222158326043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,3584,0.019487111104859244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,7168,0.031884445084465876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,2048,0.01685688893000285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,51200,0.10967377821604411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,7168,0.011317333413494958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,3584,0.27534577581617564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,10240,0.00832711077398724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,10240,0.010121777653694153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,2560,0.13591644499037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,8192,0.5564817852444118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,65536,0.15040000279744467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,8192,0.014663111832406787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,16384,0.33521244260999894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,256,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,1024,0.0074551113777690465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,768,0.010551999840471478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,12288,0.057440890206231006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,128,0.004217777815130022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,2560,0.006035555568006303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,12288,0.04856355653868782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,1024,0.011874666644467248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,3072,0.008577778107590145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,51200,3.5253492991129556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,5120,0.012156444291273752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,6144,1.0238889058430989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,12288,0.03428711162673102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,5120,0.02880711025661892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,4096,0.030881779061423406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,4096,0.013283555706342062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,768,0.01260444439119763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,1536,0.009218666288587783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,3072,0.0069120000633928515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,3072,0.010950222611427307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,8192,0.13894577821095785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,2048,0.010151111417346531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,512,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,512,0.0032924444725116095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,32,0.004102222207519743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,12288,0.02031822171476152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,1536,0.02399377855989668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,10240,0.027394667267799377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,1024,0.021720000439220007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,1536,0.0158906661801868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,51200,0.33228177494472927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,768,0.022472888231277466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,65536,0.27766667471991646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,2048,0.007149332927332983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,16384,0.3116542233361138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,7168,1.6951893700493708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,256,0.06934666633605957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,256,0.006425777657164468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,2560,0.014506666196717156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,3584,0.577423095703125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,65536,2.8475155300564237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,12288,0.0177208897140291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,1536,0.05901866488986545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,51200,0.15633422798580593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,1536,0.05644889010323418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,1024,0.02637422250376807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,6144,0.07312089204788208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,2048,0.009655999640623728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,512,0.04401422209209866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,6144,0.03188711073663499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,8192,0.012056888805495368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,3584,0.01751733322938283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,6144,0.10925155878067017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,3584,2.619071112738715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,1536,0.010533333652549319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,3072,0.09065600236256917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,10240,0.17211555110083687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,1024,0.021349334054523047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,1024,0.04642577634917366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,768,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,2048,0.00630666646692488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,8192,0.016548444827397663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,8192,0.26296178499857586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,4096,0.00977955593003167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,2560,0.006116444451941385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,6144,0.011399110986126794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,7168,0.19274666574266222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,1536,0.005858666780922148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,6144,0.42270488209194607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,128,0.025639111797014873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,8192,0.014289778139856128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,7168,0.24434844652811685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,512,0.0069448889957533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,2560,0.008309333688682979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,10240,0.6357110871209039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,2560,0.04800800151295132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,128,0.008745777938101027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,8192,0.010144000252087912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,4096,0.030118223693635728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,512,0.004576000074545543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,10240,0.0317191117339664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,32,0.004729777574539185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,7168,0.011054221954610614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,256,0.005410666680998272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,2048,0.01603911154799991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,12288,0.6198426882425944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,3584,0.0952862236234877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,512,0.005773333211739858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,6144,0.05930578046374851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,7168,0.037110222710503474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,5120,2.846474753485786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,10240,0.09189244773652817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,5120,0.012006222373909421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,32,0.0026551110463009942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,10240,0.023810666468408372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,8192,0.1509351068072849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,64,0.0027840000887711844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,10240,0.06708355744679768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,4096,0.017439110411538016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,65536,0.14940178394317627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,2560,0.1048426694340176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,16384,0.7100364367167155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,2560,0.12007021903991699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,512,0.005568000177542369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,64,0.0075733330514695905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,10240,0.06900088654624091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,3072,0.009444444543785518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,51200,8.315129597981771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,8192,0.03171199891302321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,8192,0.15799200534820557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,64,0.0025048889219760895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,10240,0.0314684444003635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,1024,0.15916889243655735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,2048,0.012706666356987424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,8192,0.050788445605172046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,12288,0.7198844485812717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,1536,0.007287999822033777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,4096,0.1445360051261054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,7168,0.02014044423898061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,768,0.011688888900809817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,51200,0.14253777927822536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,65536,21.00438944498698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,8192,0.0899511112107171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,512,0.19104088677300346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,6144,0.07459733221266004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,5120,0.011328000161382886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,7168,0.02318222158484989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,1024,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,12288,0.024698666400379602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,12288,0.03712177938885159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,512,0.008025777836640676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,256,0.08826133277681138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,51200,0.48856268988715273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,3072,0.1155173381169637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,12288,0.017501332693629794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,65536,0.03549066517088149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,51200,0.08508977625105117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,3584,0.00962488849957784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,1536,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,256,0.04341244366433886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,64,0.008183110919263627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,128,0.009352000223265754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,128,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,3584,0.12831555472479925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,1536,0.009108444054921469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,7168,0.01073333372672399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,3072,0.013194666968451606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,1024,0.020654223031467862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,2048,0.07758133278952704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,2560,0.006591111007663939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,64,0.00702133360836241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,256,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,65536,0.060376001728905566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,8192,0.3998382091522217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,12288,0.037941333320405744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,65536,4.769127315945096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,5120,0.09102221992280747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,3072,0.1480248901579115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,8192,0.26395376523335773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,256,0.007446222007274628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,5120,0.01718222267097897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,1024,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,256,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,2048,0.042062222957611084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,16384,0.14285155137379965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,5120,0.012183999849690331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,3584,0.06742044289906819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,12288,0.39046843846638996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,65536,0.23414399888780382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,3072,0.1119591130150689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,4096,0.1587617794672648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,4096,0.018742221924993727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,128,0.006537777682145436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,12288,0.01677955521477593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,7168,0.01647111111217075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,4096,0.026586666703224182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,3072,0.006079999936951532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,256,0.00498577786816491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,2048,0.029462221595976088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,1536,0.011243555280897351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,51200,0.0955386691623264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,1024,0.019316444794336956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,3584,0.0724675522910224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,1024,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,3584,0.014899555179807874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,8192,0.03571022219128079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,1536,0.006952889263629913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,1536,0.021009777983029682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,512,0.014134221606784396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,1024,0.14872977468702528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,16384,0.840578662024604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,2048,0.014966222974989148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,12288,0.8583013216654459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,3584,0.9927244186401367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,5120,0.09093599849277073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,12288,0.04539466566509671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,768,0.11220978366004096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,256,0.0037582222786214617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,16384,0.0660746693611145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,65536,11.309960259331596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,3072,0.009001777403884465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,2048,0.010967111421955956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,128,0.006000888844331105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,4096,0.009111999637550777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,6144,0.007128888534175024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,1024,0.004548444516128964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,1536,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,256,0.03071999881002638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,256,0.010058666268984476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,32,0.02530044482813941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,128,0.005470222069157495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,10240,0.0754035578833686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,3072,0.006552000012662675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,10240,4.893463134765625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,3072,0.03680088784959581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,3072,0.0215022216240565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,7168,0.05763644642300076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,1536,0.06663644313812256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,51200,0.19914844301011828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,3072,0.029318223396937054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,51200,0.052163554562462695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,4096,0.017017778423097398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,4096,0.013295110729005603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,256,0.005118222286303838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,6144,0.02495555579662323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,10240,0.058677335580190025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,1024,0.06959466801749335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,51200,0.6750151316324869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,16384,0.0500222212738461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,3072,0.19274132781558564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,10240,0.026513778501086768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,51200,4.393358018663195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,1536,0.029081778393851385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,65536,0.3147031201256646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,3072,0.012067555553383298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,4096,0.059099555015563965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,16384,0.14050489001803929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,1024,0.06356000237994723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,1024,0.00424888895617591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,3072,0.13814577791425917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,12288,0.4170080025990804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,7168,0.0671280026435852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,6144,0.09363377756542629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,32,0.0032622222271230486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,768,0.058892442120446094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,768,0.004260444392760594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,51200,0.7898266580369738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,7168,0.014225777652528552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,1536,0.16803110970391166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,3072,0.0373671121067471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,2560,0.00609688874748018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,6144,0.10086666875415379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,3584,0.5134008725484213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,4096,0.24208799997965494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,128,0.0034959999223550162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,3584,0.11614221996731228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,32,0.03659644391801622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,6144,0.010192888478438059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,4096,0.019527110788557265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,1024,0.04616533385382759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,2048,0.41347821553548175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,10240,0.015783111254374187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,256,0.007140444384680853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,1024,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,7168,0.023826667004161414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,32,0.006776888751321369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,2560,0.12669066588083902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,4096,0.13975289132859972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,12288,0.03887733485963609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,3072,0.12043733066982693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,4096,0.009338666995366415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,10240,0.08773689137564765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,16384,0.021346666746669348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,2048,0.10812889205084907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,1536,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,2048,0.008799110849698385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,512,0.007941332956155142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,2560,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,1024,0.025775111383861963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,12288,0.026867555247412786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,5120,0.03305155701107449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,512,0.008286221987671321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,12288,0.02330400049686432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,1024,0.0037600000699361167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,1536,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,256,0.0029964444951878656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,3072,0.15867555141448975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,16384,0.0217884447839525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,5120,0.017077333397335477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,1024,0.006924444602595435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,3584,0.07586399714152019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,51200,0.6664106580946181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,128,0.005059555586841372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,2560,0.01293244461218516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,32,0.0026480000880029467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,2048,0.09890577528211807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,8192,0.24156265788608125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,5120,0.028544889556037054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,4096,0.01205599970287747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,10240,0.013096888860066732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,3584,0.016714667280515034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,512,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,8192,0.3889937665727403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,16384,0.049698667393790356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,512,0.014410666293568082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,4096,0.07913066943486531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,8192,0.03779911001523336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,10240,0.029495110114415485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,2560,0.9401306576199002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,4096,0.012221333053376941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,4096,0.027136888768937852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,3072,0.008934222161769867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,7168,0.06457244687610202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,51200,6.136573367648655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,768,0.01552888916598426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,256,0.009137777818573846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,16384,0.04684799909591675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,7168,0.05355911122428047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,7168,3.697784847683377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,51200,0.7874577840169271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,3072,0.0698382192187839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,4096,0.012451555165979596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,7168,0.0476435555352105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,128,0.03235110971662734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,51200,0.019546666079097323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,1024,0.02104355560408698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,3072,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,5120,3.5816968282063804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,1536,0.005984000033802456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,8192,0.026983110441101923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,3072,0.12426933977339004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,65536,0.09513422515657212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,6144,0.01408088869518704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,1024,0.025322667426533167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,64,0.005440000030729506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,3072,0.006680000159475539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,2048,0.019401777121755812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,1536,0.006320000108745363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,32,0.03428533342149522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,2048,0.00980888886584176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,4096,0.4030631118350559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,128,0.0070088886552386815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,128,0.005727111051479976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,64,0.003351111171974076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,7168,0.22749333911471895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,10240,0.10710666577021281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,5120,0.03377244538731045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,8192,0.3331884543100993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,51200,2.1591813829210067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,1536,0.6916826566060384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,7168,0.016785777277416654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,2560,0.01812977757718828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,128,0.032030221488740705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,768,0.009469333622190688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,8192,1.22088623046875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,2048,0.008774222599135505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,1536,0.039135999149746366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,51200,0.096325331264072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,3072,0.01310488912794325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,3584,0.07794666952557035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,51200,0.24328798717922637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,1536,0.059021333853403725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,12288,0.0776488913430108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,51200,0.11840177906884088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,2048,0.044116444057888456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,32,0.0038497779104444715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,2048,0.011694221860832639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,12288,0.028047111299302842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,256,0.013510222236315409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,4096,0.01255555533700519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,8192,0.014328888720936246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,3072,0.11941956149207221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,2560,0.007682667010360294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,1536,0.013289777768982781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,2560,0.00888355573018392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,1024,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,1536,0.005617777920431561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,2048,0.007267555428875818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,1536,0.006723555425802867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,64,0.04533333248562283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,768,0.006531555619504716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,512,0.05089155501789517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,768,0.0037875556283526947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,32,0.004490666505363253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,4096,0.06327644321653578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,51200,32.41602240668403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,7168,0.01887022289964888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,2048,0.03965511255794101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,2560,0.027254222167862788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,32,0.013778666655222574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,2560,0.016920889417330425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,64,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,32,0.002573333266708586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,2048,0.016075554821226332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,1536,0.006351111249791251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,65536,0.696107546488444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,1024,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,7168,0.031382222970326744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,8192,0.027767111857732136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,65536,0.040672888358434044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,6144,0.01996977792845832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,1536,0.022504000200165644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,10240,0.8519004185994467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,32,0.0058417779703934984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,6144,0.13314489523569742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,256,0.00416711096962293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,5120,0.009305777649084726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,5120,0.02772444486618042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,16384,0.024676443801985845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,10240,0.01384622189733717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,512,0.013131555583741931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,1024,0.026732444763183594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,256,0.005064888960785336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,65536,0.06953510973188612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,512,0.0032506665835777917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,7168,0.2137706677118937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,32,0.016940444707870483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,65536,0.08050488763385348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,2048,0.10740177498923408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,5120,0.022729777627521094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,65536,0.06284266710281372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,4096,0.07597955730226305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,8192,0.019852444529533386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,256,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,16384,0.04500800040033129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,6144,0.012986666626400419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,10240,0.0251857770813836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,6144,0.06776533524195354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,128,0.01329422245422999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,3584,0.019320888651741874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,65536,1.1490373611450195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,512,0.005510222166776657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,1024,0.03604355454444885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,12288,0.006603555546866522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,5120,0.028375999795065984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,5120,0.20005510913001168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,8192,0.3688257800208197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,12288,0.04728088776270548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,12288,0.32147733370463055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,5120,0.04818755719396803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,2048,0.008886222210195329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,4096,0.007904888855086433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,7168,0.016553777787420485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,16384,0.023834667272037927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,10240,0.01568000018596649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,5120,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,128,0.0028328889360030494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,2048,0.006187555690606435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,512,0.005408888889683618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,3072,0.08371111419465806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,7168,0.19624443848927817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,51200,3.810327317979601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,1024,0.009881777895821465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,16384,0.07035644186867608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,32,0.006351999938488007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,2560,0.00870222184393141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,3072,0.014193778236707052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,512,0.004146666576464971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,12288,0.12512711683909097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,16384,1.8251635233561199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,2048,0.019018666611777413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,8192,0.060180445512135826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,1024,0.01612444387541877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,1536,0.04403733213742574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,8192,0.025123554799291823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,3072,0.07707377937104967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,512,0.004675555560323927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,3072,0.014170666535695394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,256,0.0034844444857703317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,16384,1.844940397474501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,7168,0.020560888780487906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,2560,0.009301332963837517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,65536,0.6754177941216363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,768,0.010017777482668558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,6144,0.08604622549480861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,5120,0.10684888892703587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,7168,0.08585244417190552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,10240,0.014912888407707214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,5120,0.03834755553139581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,51200,0.0952471097310384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,2048,0.01353422221210268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,64,0.032474666833877563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,3584,0.06687111324734159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,8192,0.031110223796632554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,3584,0.020801777640978496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,4096,2.256490707397461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,2560,0.009303111169073317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,256,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,12288,0.03208533260557387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,65536,1.224281734890408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,3584,0.027618666489919026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,5120,0.02875022093454997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,8192,0.14171732796563044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,32,0.0070062221752272705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,16384,0.01833422150876787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,512,0.005285333428117964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,65536,0.9028737809922961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,10240,0.04461955693033007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,32,0.003055111194650332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,128,0.03399466805987888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,1024,0.01852799952030182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,256,0.015443555182880826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,65536,0.035679999325010515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,3072,0.007780444290902879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,64,0.019987555013762582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,10240,0.07320000065697564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,2048,0.01947733263174693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,512,0.0033946666452619764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,8192,0.016522667474216886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,65536,0.41974666383531356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,10240,0.05801510810852051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,6144,0.19482221868303087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,512,0.012792888614866467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,2560,0.1570684380001492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,1536,0.006514666808976068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,12288,0.044898665613598294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,256,0.005509333478079901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,4096,0.012246222131782107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,5120,0.07746844159232245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,7168,0.10356977913114761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,2560,0.00737955586777793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,768,0.03309244579739041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,128,0.007416889071464539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,65536,0.060956445005204946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,32,0.023546665906906128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,10240,0.19166221883561876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,256,0.008450667063395182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,7168,0.6272115707397461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,6144,0.011861333416567909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,128,0.006077333456940121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,256,0.004416888786686791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,16384,0.9159946441650391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,768,0.014184888866212634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,3584,0.010833777487277985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,256,0.007890666524569193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,10240,0.01351199961370892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,8192,0.024153777294688757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,256,0.005184000151024925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,1024,0.010988444089889526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,2048,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,3072,0.021547555923461914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,2560,0.02297244469324748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,7168,0.007879999776681265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,12288,0.016807110773192514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,51200,2.561855951944987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,1024,0.10982311434215969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,12288,4.409114837646484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,65536,2.49362670050727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,3072,0.04176622298028734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,2560,0.37586310174730087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,3584,1.4458497365315754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,512,0.019973332683245342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,4096,0.4310035440656874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,10240,0.3974177837371826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,2560,0.1338640054066976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,7168,0.5712862014770508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,1536,0.014383110735151501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,5120,0.030530667967266504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,10240,0.016179554992251925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,8192,0.5531537797715929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,128,0.005418666534953647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,6144,0.06166488594479031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,128,0.0038986665507157645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,3584,0.052878220876057945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,10240,0.2756284342871772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,16384,0.07831999990675184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,32,0.005135111096832487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,768,0.016750221451123554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,3072,0.057533330387539335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,768,0.006512889017661412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,3584,0.025768889321221247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,5120,0.27549955579969615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,128,0.008191111187140146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,256,0.02896889050801595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,1536,0.01262399968173769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,2048,0.0058346668051348785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,1024,0.008632889224423302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,128,0.0717395544052124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,2048,0.012126222252845764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,64,0.003444444388151169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,256,0.017582222819328308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,4096,0.01033866653839747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,5120,0.028198222319285076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,7168,0.013945778210957846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,6144,0.05760444535149468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,4096,0.009506666825877296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,3072,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,3072,0.0872231125831604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,1536,0.01592088904645708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,1024,0.009532444179058075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,768,0.021868444151348535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,3584,0.022260444031821355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,8192,0.797361797756619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,256,0.016766221986876596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,768,0.0041697778635554844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,128,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,2560,0.0247715562582016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,64,0.2915662129720052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,512,0.012295111185974546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,64,0.003043555551105075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,10240,0.08769777748319839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,12288,0.01418133411142561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,12288,0.20258577664693198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,8192,0.014312000738249885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,51200,0.08077066474490695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,256,0.02211644417709774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,128,0.008402666283978356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,65536,0.06839199860890706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,2048,0.006735999965005451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,256,0.02095288866096073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,128,0.004834666848182678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,8192,0.03419466813405355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,6144,0.009000889129108852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,64,0.02019644445843167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,32,0.002534222271707323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,8192,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,65536,0.10967199669943915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,512,0.015047111445003085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,128,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,2048,0.00887822194231881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,10240,0.016898666818936665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,65536,1.6748026741875544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,3584,0.10713777939478557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,5120,0.9459448920355903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,256,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,8192,0.017460443907313876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,2048,0.010690666735172272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,12288,0.028098666005664404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,2048,0.036743998527526855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,16384,0.026199999782774184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,256,0.003432000055909157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,1536,0.007712889048788283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,6144,0.009641778137948778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,65536,0.3724275430043538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,8192,5.730623881022136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,12288,4.840399000379774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,3072,0.25999556647406685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,256,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,64,0.0038604442444112566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,65536,0.07220977544784546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,16384,0.07951733138826159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,16384,0.7641448974609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,512,0.01686755485004849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,32,0.014369777507252164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,3584,0.04806133442454868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,16384,0.0318524440129598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,2560,0.009138666921191746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,512,0.004553777890072929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,256,0.01975022256374359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,3072,0.05901066462198893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,10240,0.4407866795857747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,1536,0.0047493332790003884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,6144,0.04997244477272034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,512,0.010770666930410596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,65536,8.590713500976562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,65536,0.40545066197713214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,8192,0.1629048850801256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,128,0.018930666976504855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,1024,0.09814488887786865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,2560,0.13454222679138184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,256,0.002965333354141977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,3072,0.013302221894264221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,2048,0.007119110888904995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,2048,0.12117244137658013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,1536,0.014448889427714877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,1536,0.0047360000511010485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,10240,0.014169777433077494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,16384,0.037716444995668195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,65536,0.11194045013851589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,65536,2.9912079705132375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,16384,14.521369086371529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,12288,0.11321422788831924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,256,0.02532088922129737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,4096,0.023638221952650283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,32,0.0031226666437255014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,16384,0.38860355483161074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,65536,1.7417866388956706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,32,0.00509333320789867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,3584,0.012316444681750404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,512,0.008063111040327284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,5120,0.022646221849653456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,51200,0.37636976771884495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,65536,0.7884080145094129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,12288,0.07932888799243502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,2560,0.12611377239227295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,768,0.0063475556671619415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,10240,0.03552444444762336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,3072,0.020607110526826646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,2560,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,6144,0.03500266538725959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,2048,0.007059555914666917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,7168,0.009702222214804756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,16384,0.04647199975119697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,3072,0.0064275554484791225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,768,0.016604445046848722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,10240,0.02749066717094845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,768,0.010899555351999072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,1536,0.08027378055784437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,7168,0.04799555407630073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,16384,0.03257600135273404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,16384,0.03773066732618544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,64,0.003412444351447953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,6144,0.030513776673210993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,8192,0.3416124449835883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,16384,0.08455822202894424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,4096,0.010303110712104374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,16384,0.16436621877882215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,3584,0.011615110768212212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,2048,0.006287999865081575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,10240,0.017186666528383892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,16384,0.025800888737042744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,16384,0.055551999145083956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,2560,1.6146266725328233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,6144,0.01278133359220293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,16384,0.5302684572007921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,768,0.00591466658645206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,64,0.007344889144102733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,64,0.009396444592210982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,64,0.0188391109307607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,12288,0.24190754360622832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,256,0.007129777636792924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,16384,0.03207644489076402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,5120,0.017281777328915067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,16384,0.09133778015772502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,128,0.0034106667670938703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,12288,0.019363555643293593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,65536,0.014542222023010254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,12288,0.10204711225297715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,768,0.024373332659403484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,1536,0.007295110987292395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,3072,0.04402933186954922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,3072,0.027111111415757075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,4096,0.018019555343521964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,65536,33.817606608072914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,2048,0.02272355556488037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,4096,0.021015110943052504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,512,0.003399111123548614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,1024,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,7168,0.26196355289883083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,1536,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,5120,0.7107733090718588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,1024,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,768,0.00464088883664873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,1536,0.006242666807439592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,4096,0.00923911150958803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,51200,1.7272684309217665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,2560,0.0092604441775216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,256,0.003516444315512975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,512,0.004370666626426908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,12288,0.09942844178941514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,3584,0.02017244365480211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,8192,0.01418133411142561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,32,0.0031617778456873367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,1024,0.004344888859324985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,65536,0.3681928846571181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,768,0.010670222342014313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,65536,0.07096799876954821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,10240,0.06790666447745429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,6144,0.09733600086636013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,51200,0.050047109524408974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,2560,0.012692444854312472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,1536,0.03177333209249709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,4096,0.016703110602166917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,16384,0.528332445356581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,51200,0.7337155342102051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,2560,0.00942666663063897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,12288,0.16268977853986952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,768,0.04615111152331034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,5120,0.008402666283978356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,3584,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,2560,0.10579466819763184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,3072,0.0211173329088423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,1536,0.01592266725169288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,2048,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,2048,0.005614222337802251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,512,0.005899555567238066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,64,0.0027671110712819626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,2560,0.01534755527973175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,512,0.003107555624511507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,768,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,768,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,512,0.006510222123728857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,5120,0.011188444164064197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,1536,0.03358666764365302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,6144,0.44152177704705137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,4096,6.74386469523112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,512,0.008946666287051307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,4096,0.07960000303056505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,5120,0.027472888429959614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,1536,0.0074346669846110875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,256,0.04981599913703071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,65536,1.9287136925591364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,12288,0.18171200487348768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,16384,0.021272889441914026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,256,0.010236444572607676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,32,0.01696711116366916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,10240,1.343563503689236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,768,0.005048000150256687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,512,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,1536,0.011330666641394297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,2560,0.03733777668741014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,1536,0.02188355558448368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,16384,0.7589573330349393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,6144,0.22607111930847168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,64,0.0026542221506436667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,1024,1.6989564895629883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,512,0.014126222994592456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,1536,0.015468445089128284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,1024,0.062091555860307485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,128,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,10240,0.22818220986260307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,1024,0.2616177664862739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,64,0.0032879999942249725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,32,0.1457351048787435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,65536,0.1178257730272081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,256,0.008942222429646386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,7168,0.0063235556913746735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,65536,0.14937867058648005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,65536,0.1539751158820258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,3072,0.1025351087252299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,1536,0.0063893331421746155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,512,0.0032906666811969546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,5120,0.013289777768982781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,512,0.0033253334048721525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,4096,0.00637511122557852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,3072,0.006423111177153057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,2048,0.01997066703107622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,1536,0.5100764698452419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,12288,0.015214223000738354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,4096,1.55493958791097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,5120,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,51200,2.255167007446289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,2560,0.7203395631578234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,768,0.007199111084143321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,6144,0.16439821985032824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,512,0.0064479998416370815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,3072,0.009191999832789103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,768,0.0052968888647026485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,64,0.007843555675612556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,512,0.0075048887067370946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,512,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,3072,0.023743111226293776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,8192,0.032602667808532715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,2048,0.02128355536195967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,7168,0.011504888534545898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,7168,0.03374933203061422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,16384,0.23871999316745332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,256,0.0029537777105967202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,3072,0.08239733510547213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,12288,0.02109244465827942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,2560,0.02693866689999898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,3072,0.02187999917401208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,2048,0.7416506873236762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,65536,0.05702933337953356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,128,0.0036915557252036203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,16384,2.388363520304362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,3072,0.09307377868228489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,10240,0.0557804438802931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,4096,0.0613973339398702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,10240,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,64,0.009857777920034196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,32,0.15822488731808132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,128,0.007761778102980719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,5120,0.08770310878753662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,5120,0.013995556367768182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,6144,0.02902844548225403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,2048,0.006618666566080517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,2048,0.41215377383761936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,4096,0.06578399737675984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,8192,0.03096533483929104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,16384,0.33386045032077366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,32,0.00332622230052948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,3584,0.02553422252337138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,1536,0.017148445049921673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,10240,0.8234124183654785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,10240,0.016362667083740234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,12288,0.07059022453096178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,1024,0.5190569029914008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,4096,0.053660445743136935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,2048,0.013835555149449242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,512,0.005496888938877318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,256,0.015413333972295126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,3072,0.009508444203270806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,8192,0.03272977802488539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,10240,0.17957511213090685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,256,0.0029493332323100832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,51200,0.6631973584493002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,2560,0.02053333322207133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,256,0.0036968886852264404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,5120,0.013088888592190213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,768,0.0138586668504609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,3072,0.014481777946154276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,4096,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,128,0.004910222358173794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,768,0.01019200020366245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,10240,0.013638222383128272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,2560,0.045576890309651695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,4096,0.00869333330127928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,768,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,16384,0.08749510844548543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,128,0.11640000343322754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,512,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,768,0.07456533114115398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,12288,0.048100445005628795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,3584,0.020852444900406733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,3072,0.01904533306757609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,6144,0.008972444468074376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,256,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,6144,0.013399110900031196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,256,0.007598222129874759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,12288,0.09705244170294867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,6144,0.01181333346499337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,64,0.006114666660626729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,51200,0.015545777148670621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,32,0.03847644395298428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,65536,0.1363555590311686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,12288,0.017913776967260573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,5120,0.00960799968904919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,768,0.003532444437344869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,1024,0.0771155556042989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,3584,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,6144,0.08685955736372207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,65536,3.5166977776421438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,3072,0.01161688897344801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,2048,0.006688888702127669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,2048,0.005783111270931032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,8192,0.025813332862324182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,2048,0.011614222493436603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,16384,0.45262575149536133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,51200,0.055323554409874805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,65536,1.4580337736341689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,128,0.007429333196745978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,8192,0.9713786443074545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,51200,0.39244355095757383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,1024,0.006445333361625671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,512,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,768,0.01402844488620758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,512,0.0063386667105886675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,16384,6.044950273301866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,51200,0.880075560675727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,7168,0.5503644413418239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,2560,0.012155555188655853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,512,0.015879111157523263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,256,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,3072,0.007627555893527136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,3584,0.0733955568737454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,2560,0.010409778190983666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,8192,0.09424177805582683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,7168,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,16384,0.04677777820163303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,4096,0.011169777976142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,2560,0.00647288892004225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,128,0.005035555611054103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,512,0.006769777586062749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,1536,0.006576889091067844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,12288,0.8113608890109592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,4096,0.01666844387849172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,768,0.007343110938866933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,7168,0.0736959973971049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,128,0.002809777855873108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,5120,0.009876444107956356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,3072,0.007890666524569193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,12288,0.01924533314175076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,2560,0.021280889709790547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,768,0.0038382220599386427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,12288,0.3903884357876248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,65536,23.33099873860677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,2560,0.037943111525641546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,7168,1.1505582597520616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,12288,0.022413333257039387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,51200,0.2015768951839871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,2048,0.011443555355072021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,16384,0.39972533120049375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,1024,0.009025777379671732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,512,0.011361777782440186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,64,0.01088444474670622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,51200,0.10818044344584148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,65536,0.06600977977116902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,51200,0.23183732562594941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,3072,0.019524445136388142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,5120,0.015505777464972602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,256,0.019800000720553927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,6144,0.22072800000508627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,10240,0.022248889009157818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,128,0.00434666665063964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,3584,0.23839465777079263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,3584,0.1422399944729275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,2048,0.11710577540927464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,768,0.01440177857875824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,256,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,16384,0.018798222144444782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,256,0.004719999929269155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,7168,0.1311075554953681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,64,0.008936888641781276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,64,0.05958400170008341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,16384,0.32096534305148655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,3584,0.2169626686308119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,6144,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,65536,0.29597245322333443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,256,0.0047591109242704184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,65536,2.0370676252577042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,4096,0.014801777071423002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,4096,0.014945778581831189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,51200,0.20272355609469941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,256,0.01590755581855774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,4096,0.02249777813752492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,256,0.0029724445194005966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,1024,0.008783111141787635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,12288,0.07564089033338758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,2560,0.026084444589085046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,2560,0.8375137646993002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,65536,0.4035075505574544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,512,0.005230222311284807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,7168,0.019256000717480976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,1536,0.0047155556579430895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,7168,0.011586666935020022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,7168,0.18413866890801323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,512,0.017680888374646504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,128,0.005247999810510212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,2560,0.013998222019937305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,16384,0.3765386740366618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,4096,0.011705777711338468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,51200,0.06465333037906222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,6144,0.016036444240146212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,10240,0.3032906585269504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,512,0.016028443972269695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,12288,0.044181333647833936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,4096,0.3200613392723931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,65536,0.05218133330345154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,3584,0.007263999846246507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,1536,0.005737777799367905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,8192,0.025601777765485976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,7168,0.09790488746431138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,2048,0.02770400047302246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,65536,1.9209022521972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,8192,0.030566222137875024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,51200,0.15275377697414821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,12288,0.821166197458903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,5120,0.16319110658433703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,32,0.007691555553012424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,3072,0.018392889036072623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,51200,0.26914310455322266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,1024,0.012540444731712341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,768,0.008365333080291748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,16384,0.058443552917904325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,5120,0.1357768906487359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,512,0.025295999315049913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,32,0.022520888182852004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,512,0.006021333237489064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,1536,0.011086222198274402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,128,0.005460444423887465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,512,0.00332622230052948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,2560,0.09598044554392497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,7168,0.015097777048746744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,512,0.005541333307822545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,2048,0.010545777777830759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,65536,0.19040355417463514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,2560,0.008815999660227034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,65536,0.310806221432156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,5120,0.018926221463415358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,3584,0.01998222205373976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,768,0.014363555444611443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,5120,0.43281955189175075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,16384,0.019461333751678467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,1024,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,1536,0.5048791037665473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,12288,0.018770666586028207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,768,0.03983733389112685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,1024,0.008347555167145198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,12288,0.31643109851413304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,6144,0.16151555379231772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,768,0.004782222211360931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,6144,0.03028622269630432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,2560,0.015409777561823526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,7168,0.02237244447072347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,1536,0.08735022279951309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,32,0.011543999943468304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,4096,0.02020355562369029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,5120,0.024114666713608637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,1536,0.009183111290136972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,128,0.002838222309947014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,3584,0.1277057727177938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,1536,0.006216889040337668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,1024,0.05474133292833964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,2048,0.09053688579135472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,65536,0.06917511092291938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,5120,0.00888177752494812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,5120,0.06627644432915582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,65536,1.3509706921047633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,12288,0.07007022036446466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,5120,0.043266667260064014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,51200,0.02537244392765893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,10240,0.038937777280807495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,10240,0.3013333214653863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,6144,0.012375999655988483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,10240,0.016538666354285348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,768,0.025469332933425903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,4096,0.023930667175187006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,1024,0.026879999372694228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,12288,0.009675555759006077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,2560,0.020169778002632987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,128,0.0033075554917256036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,8192,0.7024808989630805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,4096,0.022443556123309668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,16384,0.9799039628770617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,768,0.007615110940403408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,1536,0.00665155549844106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,128,0.0206986665725708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,51200,0.09580533372031318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,4096,0.012495111260149213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,8192,0.2683013280232747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,16384,0.05190844337145487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,5120,0.02440888848569658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,10240,0.38201067182752824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,2560,0.029197331931855943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,8192,0.030960887670516968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,3584,0.14063378175099692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,51200,0.16986489295959473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,2048,0.06493777698940702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,4096,0.010640000303586325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,1536,0.07437422540452746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,128,0.004643555730581284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,256,0.0033733333564466904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,8192,0.720762676662869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,3584,0.0469084448284573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,65536,0.11803466743893093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,7168,0.02472711106141408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,2560,0.061956445376078285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,128,0.0033528889632887314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,65536,0.529587533738878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,12288,0.1022462248802185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,3584,0.032315555546018816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,3584,0.3498666551378038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,12288,0.017276444368892245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,1536,0.004881777697139316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,6144,0.014526221487257214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,1536,0.015476443701320224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,2560,0.2880435519748264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,3584,0.014875556031862894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,2560,0.08175022072262235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,4096,0.11354755030737983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,256,0.0029173334025674393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,512,0.01551999979548984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,10240,0.010197333163685268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,5120,0.02604266670015123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,4096,0.017498667041460674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,1024,0.044504887527889676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,768,0.003583111282851961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,128,0.0047324444684717385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,768,0.004026666697528628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,2048,0.03051911128891839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,64,0.0026266665922270883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,6144,0.1188711060418023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,16384,0.6915546523200141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,1024,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,6144,0.05270222160551283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,768,0.008220444122950235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,3584,0.2482586701711019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,7168,0.20437244574228922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,12288,0.08899999989403619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,3072,0.008577778107590145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,65536,0.09460088941786025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,4096,0.01978400018480089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,1536,0.060432891050974526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,16384,0.02492800023820665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,12288,7.3636686537000875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,2048,0.01163200040658315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,4096,0.5342924329969618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,12288,0.3370053238338894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,128,0.004892444445027245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,2048,0.0776453349325392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,2048,0.009353777600659264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,12288,0.18371821774376762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,64,0.02455733385350969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,51200,8.987788730197483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,10240,0.1561866601308187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,1024,0.04306133257018196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,7168,0.20972088972727457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,256,0.007153777612580194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,2560,0.27926932440863717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,3584,0.009753777749008602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,65536,0.09818577766418457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,10240,0.06047022342681885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,5120,0.36818578508165145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,256,0.011078221930397881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,4096,0.011243555280897351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,51200,0.06540622313817342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,128,0.00849333322710461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,1536,0.01204711116022534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,6144,0.11120443873935276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,12288,0.5036995675828722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,65536,0.42295466528998477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,65536,0.15881244341532388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,768,0.007754666937722101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,10240,0.09686133596632217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,6144,0.02908088763554891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,3072,0.006542222367392645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,3584,0.009495110975371467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,51200,0.7845982445610894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,5120,1.8525405459933812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,3584,0.012291555603345236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,256,0.021945777866575453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,128,0.010750222537252637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,10240,0.028959999481836956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,128,0.10811110999849106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,8192,0.01685422162214915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,8192,0.013954665925767688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,32,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,10240,0.024306666519906785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,12288,0.3030711015065511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,12288,0.02092622220516205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,16384,0.052424887816111244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,7168,0.27906844351026744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,7168,0.3605973455641005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,8192,0.30171910921732586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,64,0.0026826666047175727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,4096,0.011914666328165265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,8192,0.03350311186578538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,12288,0.09497155745824178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,16384,0.7531484497918023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,10240,0.7157768673366971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,128,0.012856889102194043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,16384,0.41603022151523167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,256,0.00608800020482805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,5120,0.023920888702074688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,1536,0.01332355539004008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,6144,0.6554169125027126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,2560,0.01811911165714264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,512,0.00701333334048589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,5120,0.13091111183166504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,12288,0.028195556667115953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,32,0.012229333321253458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,2048,0.010429333481523724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,128,0.003112000102798144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,8192,0.04810666706826952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,256,0.02608000073168013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,10240,28.14978535970052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,2560,0.016113777955373127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,51200,1.3550969229804144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,64,0.009549332989586724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,128,0.004716444346639845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,6144,0.15192177560594347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,7168,0.009157333109113906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,8192,0.007344889144102733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,1024,0.028499556912316218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,2048,0.011031111081441244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,51200,0.06417422162161933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,2048,0.19320266776614717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,16384,0.1904613309436374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,2560,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,4096,0.1462426715426975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,64,0.0026942222482628296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,7168,3.5637084113227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,16384,0.22832798957824707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,2560,0.01645866698688931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,256,0.00886044485701455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,12288,0.02358666724628872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,512,0.005959110955397288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,64,0.0027840000887711844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,1536,0.010656000011497073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,1536,0.013609777722093793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,8192,0.06558755371305677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,12288,0.02740977704524994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,10240,0.013936888840463428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,4096,0.16104443868001303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,2048,0.04406222369935778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,256,0.006484444356626934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,5120,0.01296266665061315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,65536,0.15339110957251653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,7168,1.500780423482259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,512,0.007463999920421177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,10240,0.01756533318095737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,10240,0.15450933244493273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,5120,0.011211555865075854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,1536,0.009935110807418823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,768,0.006416000011894438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,3584,0.019895111521085102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,2048,0.02273511058754391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,768,0.014782221780882942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,8192,0.2859475612640381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,2560,0.008971555365456475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,1024,0.012816888590653738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,65536,1.7309883965386286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,1536,0.03173689047495524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,3584,0.7530453470018176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,51200,0.48271555370754665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,65536,0.13882044951121011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,12288,0.31721509827507866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,16384,0.10371822118759155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,4096,0.010747555229398938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,2048,0.032368001010682844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,256,0.025441777375009324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,51200,0.011872889267073737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,5120,3.1090382470024953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,65536,0.6056995391845703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,256,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,1024,0.03994399971432156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,10240,0.17286933792961967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,2048,0.2166799969143338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,2560,0.06795199712117513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,2048,0.06917333602905273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,6144,0.014691554837756686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,5120,0.02769600020514594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,768,0.011237333218256632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,7168,0.017525333497259352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,1536,0.0079644446571668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,3584,0.010866666833559671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,768,0.016918222109476726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,2560,0.01222222215599484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,7168,0.15008444256252712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,65536,0.2468444506327311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,1536,0.009333333207501305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,3584,0.010713777608341642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,1024,0.096888886557685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,1024,0.03157155712445577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,51200,0.2259164386325412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,3072,0.015545777148670621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,16384,0.048193779256608754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,32,0.4602710935804579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,5120,0.02056711084312863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,65536,0.06827555762396918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,32,0.022683555881182354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,512,0.004801777915822135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,5120,0.025991999440722998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,4096,0.11550222502814399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,3584,0.007843555675612556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,2560,0.006581333362393909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,1024,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,7168,0.019744000501102872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,51200,1.3413964377509222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,5120,0.2074906693564521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,256,0.011415111521879831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,51200,211.57722981770834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,4096,0.023984889189402264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,512,0.00342399999499321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,5120,3.3606319427490234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,51200,0.582655111948649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,2560,0.0075297777851422625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,6144,0.012700444294346703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,128,0.0030737777964936364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,7168,0.01441422270403968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,7168,0.08869777785407172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,7168,0.0136835558546914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,65536,0.9818933275010852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,512,0.004061333421203825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,5120,0.19132888317108154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,8192,0.015821332732836407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,128,0.008485332959228093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,3584,0.014703111516104804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,10240,0.039684444665908813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,7168,0.31192265616522896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,7168,0.5421733326382107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,10240,0.04030133287111918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,7168,0.014732443624072604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,1536,0.016629333297411602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,256,0.014244443840450711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,51200,1.5872009065416124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,10240,0.04779377910825941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,16384,0.09019466903474595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,128,0.011357333097192975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,1024,0.01296266665061315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,12288,0.9740693834092883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,6144,0.05470488799942864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,12288,0.08088711235258314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,3584,0.01202933324707879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,5120,0.0785635577307807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,10240,0.0496346652507782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,6144,0.21795733769734701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,2560,0.9183297687106662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,8192,0.4940515624152289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,512,0.0033315554675128725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,6144,0.713107532925076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,16384,0.38619110319349503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,10240,0.33229955037434894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,10240,0.03498844305674235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,3072,0.10388622019026016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,2560,0.00655644428398874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,51200,0.3853377766079373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,3584,0.006792888873153263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,65536,0.8667075369093152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,32,0.008593777815500895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,8192,0.2186515596177843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,12288,0.020100444555282593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,10240,1.0962613423665364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,5120,0.011223110887739392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,512,0.004914666629499859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,1536,0.6443715625339085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,768,0.011027555498811932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,256,0.01538133290078905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,256,0.02861066659291585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,16384,0.3075804445478651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,256,0.012874667015340594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,3584,0.5809751086764866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,5120,0.032744000355402626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,32,0.007317333585686154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,512,0.006001777946949005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,8192,0.02646933330429925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,3072,0.5460160043504503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,8192,1.2424738142225478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,1024,0.007453333172533248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,12288,0.041788445578681103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,6144,0.014465777410401238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,768,0.005011555635266834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,4096,0.018047110901938546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,128,0.10683821969562107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,51200,0.48460621303982204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,768,0.003974222060706881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,32,0.014310222533014087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,6144,0.021007110675175984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,2560,0.02480799953142802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,4096,0.015411555767059326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,12288,0.17356533474392363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,2560,0.07913955714967516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,1536,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,16384,0.026178666286998328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,10240,0.23093689812554252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,32,0.005326222214433882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,4096,0.009657777845859528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,64,0.007687110867765214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,3072,0.01955111159218682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,3584,0.013023999830087027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,768,0.46918757756551105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,1024,0.21237422360314262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,2560,0.017234666479958426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,65536,0.08462844292322795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,12288,0.021143999364640977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,3072,0.09869422515233357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,8192,0.0952862236234877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,128,0.004348444441954295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,16384,0.011696889168686338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,256,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,10240,0.008319999608728621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,512,0.008464888566070134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,256,0.003563555578390757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,768,0.006774222271309958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,51200,0.11121600204043919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,5120,0.193923552831014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,10240,0.12205599413977729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,2560,0.009542221824328104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,768,0.004072000169091755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,2560,0.012352888782819113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,128,0.005394666559166378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,51200,0.06441689199871488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,128,0.02216177847650316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,64,0.03603200117746989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,65536,0.22155466344621447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,5120,0.009810667071077559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,2048,0.01574133336544037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,3584,0.016176000237464905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,16384,0.28552622265285915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,2048,0.009508444203270806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,10240,0.6654319763183594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,10240,0.42023645506964785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,51200,0.5785759819878472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,10240,0.01734044485621982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,16384,0.05942400296529134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,1536,0.03784444597032335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,4096,0.022970666488011677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,3072,0.008572444319725037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,512,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,6144,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,3072,0.10622666941748725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,3584,0.01806311143769158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,10240,0.0444160004456838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,51200,0.9133733113606771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,768,0.12302488750881618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,12288,0.28528356552124023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,1024,0.006934222247865465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,1024,0.011320888996124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,4096,0.11115111245049371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,128,0.003085333233078321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,2560,0.020406223005718656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,2048,0.005600000007285013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,65536,0.7034862306382922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,7168,0.08139910962846544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,64,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,3584,0.25813243124220103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,65536,0.9914809332953559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,10240,0.5600488980611166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,4096,1.1807830598619249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,32,0.006177777631415262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,32,0.004259555704063839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,256,0.014648889501889547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,1024,0.010911111202504901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,1024,0.0704808897442288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,768,0.027790221903059218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,12288,0.019136889113320243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,51200,0.05533511108822293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,12288,0.3173191017574734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,3584,0.2036177714665731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,65536,0.2812266614702013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,512,0.004416000097990036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,256,0.004911111046870549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,16384,1.8363634745279949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,6144,0.0493253337012397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,64,0.009153777526484596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,51200,0.4236684375339084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,256,0.0402746664153205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,4096,0.23252622286478677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,10240,0.014503111441930136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,3072,0.010020444790522257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,2048,0.19628443982866076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,3072,0.01899466746383243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,3072,0.006694222076071634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,2048,0.09172355466418797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,12288,1.0923058191935222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,256,0.005869333528810077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,2560,0.018826666805479262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,3584,0.009295110901196798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,7168,0.3026319874657525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,7168,0.7806133164299859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,2560,0.0061048890153567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,3584,0.010905777414639791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,12288,0.18278133869171143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,2560,0.02199555602338579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,3584,0.007978666987684038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,2048,0.014367111855083041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,512,0.00906577789121204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,8192,0.03248088889651828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,65536,0.3231102095709907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,6144,0.04551288816663954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,3072,0.047437333398395114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,5120,0.031918221049838595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,4096,0.006220444622966979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,256,0.003047111133734385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,1024,0.01180622229973475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,512,0.11984622478485107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,51200,0.02453244394726223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,256,0.005453333258628845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,6144,0.06281689140531752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,8192,0.5949955516391331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,12288,0.021793777743975323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,10240,0.029455112086402044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,3584,0.02735200027624766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,6144,0.006054222169849608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,128,0.0034462221794658234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,65536,0.08529777659310235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,4096,0.16026043891906738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,2048,0.02567200031545427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,256,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,512,0.01959377858373854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,4096,0.14727200402153864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,3584,0.10782221953074138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,3072,0.007686222592989604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,7168,0.005872888697518243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,768,0.004156444635656145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,3584,0.24182844161987305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,12288,0.1056542197863261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,64,0.003491555651028951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,768,0.07632711198594835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,6144,0.013437333206335703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,8192,0.007197333706749811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,768,0.06646755668852064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,1536,0.027934223413467407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,65536,0.04911999901135763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,32,0.019153777096006606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,2048,0.005969777703285217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,5120,0.20565244886610243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,128,0.011765333513418833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,6144,1.9523048400878906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,1024,0.008976000050703684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,5120,0.030156443516413372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,5120,0.08299644125832452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,512,0.012802666260136498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,128,0.010659555594126383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,768,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,6144,0.09374488724602593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,7168,0.025987555583318073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,4096,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,128,0.017251556118329365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,7168,0.00793333351612091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,3072,0.00924177798959944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,2560,0.07073511017693414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,51200,6.433064778645833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,256,0.008657777474986183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,65536,0.2574311097462972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,4096,0.04752711123890347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,128,0.003014222201373842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,4096,0.011172444456153445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,7168,0.1021511091126336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,5120,0.00980977796845966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,51200,0.16193600495656332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,64,0.002566222308410539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,8192,0.2685493363274468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,128,0.012783110969596438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,51200,0.11723555458916558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,7168,0.013408888545301227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,51200,0.2687893443637424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,1024,0.008113777471913232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,3072,0.038557334078682795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,7168,0.01717688971095615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,6144,0.010595555934641095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,768,0.0037155557009908888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,128,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,10240,0.2569119930267334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,512,0.0053004444473319584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,256,0.014157333307796054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,1024,0.008336889247099558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,128,0.005152889009979037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,128,0.021232888102531433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,4096,0.006214222146405115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,16384,0.49314313464694554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,512,0.011880888707107969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,10240,0.03943822118971083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,128,0.006053333481152852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,2560,0.007274666594134436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,65536,0.4205653402540419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,768,0.009528888596428765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,1536,0.008594666918118795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,6144,0.2112506628036499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,16384,0.02134311033619775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,64,0.010084444450007545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,2560,0.008263111114501953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,16384,0.007245333658324347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,8192,0.11062400208579169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,65536,1.5381991068522136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,256,0.022795556320084467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,128,0.14065778255462646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,2560,0.4169706768459744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,2560,0.013576889203654395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,65536,1.0058533350626628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,5120,0.060475554731157094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,10240,0.04821244544453091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,3072,0.18056266837649873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,8192,0.8644151157803006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,4096,0.6189715597364638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,2560,0.007913777397738563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,3584,0.02142133315404256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,1024,0.015800888339678448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,12288,0.023757333556811016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,3072,0.09446933534410264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,32,0.0033555556502607134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,12288,1.8362365298801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,7168,0.04643911123275757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,16384,0.2405946784549289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,4096,0.14974221918318006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,64,0.02215911116864946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,8192,0.014836443795098199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,4096,0.01720177796151903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,3584,0.00960177762640847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,65536,0.6934524642096626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,512,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,1536,0.025855110751258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,51200,12.28063710530599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,768,0.2599671151902941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,12288,0.47025330861409503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,65536,0.15895733568403456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,51200,0.03978844483693441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,5120,0.12548444006178114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,768,0.009876444107956356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,3072,0.44022311104668516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,10240,0.04263200031386482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,12288,0.019650666250122916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,256,0.0057608890864584185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,3072,0.24106844266255698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,8192,0.03127288818359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,1024,0.009621333744790819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,16384,0.2058080037434896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,1024,0.3622764481438531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,512,0.00406666679514779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,3584,0.052482667896482676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,5120,0.02387822171052297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,3584,0.00942755573325687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,512,0.0038533334930737815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,65536,0.2003493309020996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,1536,0.07135733630922106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,5120,0.16650044918060303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,16384,0.05495555533303154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,5120,0.4069048828548855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,1536,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,1536,0.0075128889746136135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,8192,0.028415110376146104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,128,0.005013333426581489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,64,0.023335999912685815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,2560,0.05185777942339579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,64,0.018935110833909776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,2048,0.01610933409796821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,3584,0.007845333053006066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,512,0.009109333157539368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,51200,0.01889600025282966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,65536,0.35504534509446883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,2560,0.015789333317014907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,1536,0.01605244477589925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,5120,0.1253582239151001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,512,0.007987555530336168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,3072,0.009670221971140968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,64,0.008063111040327284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,12288,0.05793955590989855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,8192,0.18457155757480195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,10240,0.12177599800957574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,2048,0.011901333100265928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,1024,0.034513778156704374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,1024,0.035869333479139544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,8192,0.020745777421527438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,12288,0.18006843990749785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,8192,0.38187376658121747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,32,0.01495822270711263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,10240,0.013603555659453074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,128,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,512,0.01407911048995124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,2560,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,4096,0.006846222198671765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,4096,0.042753779225879245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,8192,0.3202035427093506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,4096,0.17287466261121961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,256,0.0088764445649253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,8192,0.012079111403889127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,1536,0.030728889836205378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,10240,0.01718755563100179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,1536,0.0107342220014996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,16384,0.051022221644719444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,6144,0.22583021057976616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,768,0.011730666789743634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,512,0.024215110474162634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,6144,0.11408978038363987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,256,0.021125333176718816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,1536,0.07172266642252605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,768,0.031082666582531396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,6144,0.015284443895022074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,5120,0.0902880032857259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,128,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,5120,0.009826666778988307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,16384,0.1289422247144911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,51200,0.2957715458340115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,512,0.004378666480382283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,1024,0.04555822081036038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,32,0.0034808889031410217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,3072,0.01977511081430647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,512,0.005679111099905438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,12288,0.376251565085517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,64,0.04857777886920505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,6144,0.07154666715198092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,5120,0.07692178090413411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,512,0.05962044662899441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,51200,0.0323884454038408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,5120,0.11148533556196426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,3072,0.009838222629494136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,5120,0.13572355111440024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,6144,0.09243110815684001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,128,0.008987555901209513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,2048,3.3340738084581165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,2048,0.00535999983549118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,768,0.15461956130133733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,3584,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,3072,0.02218933403491974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,1536,0.008404444489214156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,65536,0.05747644768820869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,16384,0.02721066607369317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,256,0.00572622236278322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,768,0.029410666889614526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,2048,0.015933333171738517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,65536,0.39995021290249294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,16384,1.3500906626383464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,2048,0.33120889133877224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,64,0.009552000297440423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,5120,0.018790221876568265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,512,0.0058702222175068325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,2048,0.007824000385072496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,12288,0.03450311223665873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,51200,10.081854926215277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,1024,0.004451555510361989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,65536,0.16599110762278238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,3584,0.033668445216284856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,7168,0.3925146791670058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,128,0.04926844437917074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,2560,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,8192,0.08356089062160915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,6144,0.21740267011854383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,128,0.00490399988161193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,51200,0.6875119739108615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,12288,0.024525332782003615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,512,0.09809688727060954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,65536,0.1350017786026001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,2560,0.05836711327234904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,2048,0.007365333537260692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,768,0.008541333178679148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,12288,0.034527109728919134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,3072,0.013027555412716336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,12288,0.015023110641373528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,51200,0.04135466615358988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,256,0.007135999699433644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,2048,0.2629964351654053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,51200,0.05460888809627957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,768,0.029665778080622356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,128,0.036044445302751325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,128,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,2560,0.008421333299742805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,16384,0.033944000800450645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,256,0.17694932884640166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,51200,17.387676662868923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,3072,0.014164444473054675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,1536,0.04227822356753879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,64,0.007992000215583378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,1024,0.004671999977694618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,10240,0.5130986637539333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,2048,0.007152888509962294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,512,0.030680888228946265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,10240,0.41839912202623153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,4096,0.020319999919997323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,5120,0.016117332710160148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,8192,0.06578310992982653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,4096,0.006875555548402999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,7168,0.013880000346236758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,10240,0.01351999988158544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,10240,0.035331557194391884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,4096,0.025143999192449782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,1536,0.006353777729802662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,7168,0.056319110923343234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,10240,0.1756382253434923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,3584,0.23961512247721353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,32,0.002577777744995223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,2048,0.007483555210961237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,768,0.03978488842646281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,6144,0.3887724346584744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,5120,0.016143111719025504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,1536,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,51200,0.40614043341742617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,5120,0.02882400155067444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,12288,0.025584888127115037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,128,0.00425155543618732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,64,0.04357599880960253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,10240,0.05496266815397474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,64,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,2048,0.6117848820156521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,4096,0.023589332898457844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,768,0.004584889031118817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,2048,0.038247111770841814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,10240,1.6318818198310003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,7168,0.030880000856187608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,8192,0.025504888759719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,6144,0.029834667841593426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,65536,0.33898221121893984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,51200,23.44171142578125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,5120,0.01612622208065457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,6144,0.009418666362762451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,1024,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,6144,0.0063279999627007385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,3072,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,1024,0.025211556090248957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,128,0.00519466648499171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,2048,0.007808888951937358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,5120,3.441097683376736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,128,0.004774222357405556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,7168,0.06187466780344645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,256,0.0030417777597904205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,768,0.006082666830884085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,128,0.1739208830727471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,51200,0.15260088443756104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,51200,2.073727077907986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,32,0.3430657651689317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,65536,0.09249511029985215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,768,0.00739555557568868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,256,0.004836444639497333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,256,0.014827556080288358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,3584,0.009081777599122789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,12288,0.030376000536812678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,8192,0.062236441506279834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,64,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,3584,0.08255733384026422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,2560,0.011454222102959951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,8192,0.016813332835833233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,3072,0.010936000280910067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,256,0.0032008888406885993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,51200,0.9661457273695203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,4096,0.02182933357026842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,10240,0.014389332797792224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,2048,0.011052444577217102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,65536,0.7877795431349012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,3584,0.011583111352390714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,1536,0.2799599965413411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,2560,0.015630222029156156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,12288,3.5568817986382375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,256,0.0277173337009218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,768,0.29408711857265896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,2560,1.3609191046820746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,8192,0.011797332929240333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,10240,0.008002666963471307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,128,0.02497688929239909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,512,0.021242666575643752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,3072,0.01165422217713462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,5120,0.441177765528361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,51200,0.07187644640604655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,6144,0.08924088875452678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,51200,0.07316533062193128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,3584,0.08219733503129747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,512,0.005752888818581899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,2048,0.14303733242882624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,4096,0.30458754963344997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,128,0.14847555425431994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,7168,2.026524437798394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,12288,0.020470221837361652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,256,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,1024,0.0092284447617001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,1536,0.008618666893906063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,1536,0.06148799922731188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,1536,0.02106399999724494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,16384,0.07124533255894978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,2560,0.050316443045934044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,51200,2.1780054304334855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,3584,1.1624630822075737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,1536,0.00869777798652649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,51200,0.2682897779676649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,16384,0.036152889331181846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,7168,0.020114666885799833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,2560,0.34626399146185977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,10240,0.05176000131501091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,16384,0.03137866655985514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,12288,2.4113441043429904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,5120,0.0335440006521013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,4096,0.02509511179394192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,51200,3.6884935167100696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,7168,0.11001511414845784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,2048,0.07495911253823175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,512,0.006430222342411677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,65536,0.1814284457100762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,3072,0.009961778091059791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,32,0.0027511111564106415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,3072,0.011004444625642566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,5120,0.013905778527259827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,10240,0.01666844387849172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,1024,0.005763555566469829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,12288,0.08138400316238403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,51200,0.05331910981072319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,12288,0.09395111269421047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,768,0.004921777794758479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,3584,0.013045333325862885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,1024,0.009141333401203156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,1536,0.004629333400064045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,4096,0.04472088813781738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,16384,0.594737794664171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,51200,0.15690488285488552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,6144,0.5199324289957682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,64,0.0336791111363305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,1536,0.007054222126801808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,256,0.0041804446114434134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,16384,0.05955288807551066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,2560,0.01564800077014499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,3584,0.010079999764760336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,5120,0.008685333033402761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,3072,0.3371528784434001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,3584,0.012678222523795234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,7168,0.684300422668457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,3584,0.015067555838161044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,7168,0.01834488908449809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,2560,0.06456889046563043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,7168,0.09334755606121486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,768,0.0129439996348487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,12288,0.018907555275493197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,512,0.01609511176745097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,128,0.0047004446387290955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,64,0.04653155472543505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,16384,0.026713778575261433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,3072,0.03577955563863119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,10240,0.024400888217820063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,768,0.00370399985048506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,5120,0.2801520029703776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,5120,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,3584,0.14722843964894614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,4096,1.5130382113986547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,16384,0.012316444681750404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,65536,0.1726142168045044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,32,0.006115555349323485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,1536,0.007890666524569193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,8192,0.006504000061088138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,8192,0.2672773202260335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,51200,7.1151928371853295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,7168,0.013288000391589271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,12288,0.026446221603287592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,2560,0.017743999759356182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,8192,0.03151199883884854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,768,0.1258337762620714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,6144,0.3743155532413059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,1024,0.031531555785073176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,32,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,3072,0.013410666750537025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,12288,0.3088906606038411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,2560,0.02184711065557268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,12288,0.020115555988417733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,51200,0.045682665374543935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,3072,0.009341333475377824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,512,0.01498222185505761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,128,0.0064791109826829694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,7168,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,8192,2.8787583245171438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,2048,0.07341600126690336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,2048,0.015340444114473132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,65536,0.10073244571685791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,64,0.007244444555706448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,51200,0.16392266750335693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,128,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,12288,0.0640853312280443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,5120,0.011495999991893768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,4096,0.2731342315673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,10240,0.08166399929258558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,12288,0.023043556345833674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,7168,0.016883555385801528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,4096,0.0277511113219791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,6144,0.05923910935719808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,3072,0.021401777863502502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,64,0.005613333235184352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,2560,0.016025778320100572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,3072,0.021890666749742296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,3072,0.008579555484983657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,1536,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,2560,0.00980177770058314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,1024,0.006215111249023014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,3584,0.030518223841985066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,65536,0.040023999081717596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,2048,0.00830666638082928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,128,0.002838222309947014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,2560,0.2555475499894884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,10240,0.006073777874310811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,1536,0.03228266702757941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,256,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,51200,0.4490346908569336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,65536,0.07717421982023451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,64,0.0686017804675632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,1536,0.038069334295060896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,5120,0.00959288908375634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,2560,0.022433777650197346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,2048,0.028713779317008123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,1536,0.00738844441043006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,12288,3.3386107550727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,51200,2.2208720313178167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,16384,0.21667822202046713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,3584,0.026372444298532274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,2048,0.007330666813585494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,128,0.07736800114313762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,128,0.010188444621033138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,12288,0.029901332325405542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,1024,0.4037324322594537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,6144,0.03615555498335097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,1536,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,12288,0.08838133017222087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,256,0.0030746666921509635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,16384,0.021346666746669348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,5120,0.010970667004585266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,768,0.009591999981138442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,256,0.002973333415057924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,3584,0.022985777921146814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,3072,0.13168088595072427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,3584,0.006498666687144174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,10240,0.012251555919647217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,16384,0.2978124353620741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,12288,0.023426666855812073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,51200,0.12037777900695801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,768,0.029558221499125164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,12288,0.06480089161131117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,16384,0.02064533366097344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,256,0.011897777517636618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,8192,0.01645955608950721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,8192,4.117171393500434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,5120,0.23142843776279023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,16384,0.8144542376200358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,6144,0.05596178107791477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,51200,0.008992888861232335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,65536,0.1565342214372423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,7168,0.019494222270117868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,10240,0.09714577595392863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,5120,0.014171555638313293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,16384,0.6066862212287055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,1024,0.006189333481921091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,5120,0.31460799111260307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,51200,0.3918195565541585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,7168,0.01833066675398085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,65536,0.37906667921278214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,3584,0.18436266316307914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,2048,0.00848444468445248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,4096,0.019663110375404358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,1024,0.035655998521380954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,3072,0.007576000359323289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,16384,0.0648346675766839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,1024,0.01109688894616233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,3584,0.008080888953473832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,4096,0.15536177158355713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,512,0.004743111216359668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,8192,0.049350221951802574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,128,0.0026577777332729767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,2560,0.00794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,256,0.003105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,4096,0.01109955542617374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,32,0.0038693332009845306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,32,0.0029253332565228143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,3584,0.5412169032626682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,6144,0.764146645863851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,2560,1.2294026480780709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,51200,0.009573332965373993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,16384,0.12619377507103816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,2560,0.05703555875354343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,128,0.008529778156015608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,65536,0.21817599402533638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,3072,0.08905777666303848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,512,0.005427555491526921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,4096,0.14109599590301514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,10240,0.020259555843141343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,10240,0.019710222880045574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,32,0.025941333836979334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,2560,0.008029333419269985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,768,0.04734399914741516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,3584,0.01716800034046173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,3072,0.2833964559766981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,51200,2.0046070946587458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,768,0.0046728890803125174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,12288,0.17784178256988525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,2560,0.014545778433481852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,51200,0.17623999383714464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,65536,0.4015582137637668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,768,0.010447111394670276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,2560,0.012465777496496836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,5120,0.036023110151290894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,1536,0.019157333506478202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,8192,0.03907022211286757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,2560,0.00962222201956643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,65536,0.08056088950898913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,6144,0.13738311661614314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,7168,0.02457333273357815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,2560,0.017111111018392775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,4096,0.09149599737591213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,1536,0.010116444693671333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,7168,0.03269155489073859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,4096,0.04877600073814392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,128,0.015585778488053216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,4096,1.1946836047702365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,10240,0.11427911122639973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,65536,6.485950046115452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,256,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,4096,0.013495999905798169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,512,0.007607111500369177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,3072,0.03680977887577481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,3072,0.018467555443445843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,3072,0.00815555536084705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,1536,0.049775110350714795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,10240,0.017989334132936265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,65536,0.6730275683932834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,1536,0.012712889247470431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,51200,19.209788004557293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,12288,0.23574399948120117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,3584,0.02306844459639655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,512,0.004967999955018361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,1536,0.005665777872006099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,1024,0.010487111078368293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,65536,0.24973156717088488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,5120,0.009376889301670922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,3072,0.018762666318151686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,1536,0.01572533282968733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,512,0.296195560031467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,3072,0.015092444088723926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,2048,0.017282666431532968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,65536,0.10614933570226033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,1024,0.016045333610640634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,8192,0.6849857966105143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,128,0.003585777762863371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,768,0.007116444408893585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,5120,0.02201955517133077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,1536,1.904295179578993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,7168,0.03133066826396518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,7168,0.030589332183202107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,12288,0.025621333056026038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,51200,3.2866649627685547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,3584,0.01000355515215132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,16384,1.0665288501315646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,1024,0.11801689200931126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,65536,0.21144533157348633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,256,0.022072000635994807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,6144,3.754570855034722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,12288,0.07964889208475749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,10240,0.01239911135700014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,6144,0.01462577780087789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,8192,0.009004444711738164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,16384,0.2520835664537218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,8192,0.0063075555695427795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,768,0.016805332567956712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,51200,0.053800887531704374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,2048,0.005981333139869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,16384,6.11003918117947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,16384,0.48680798212687176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,65536,0.3203786744011773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,4096,0.011810666984981961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,2560,0.009845332966910468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,32,0.003964444415436851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,2048,0.009152000149091085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,64,0.0058764442801475525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,12288,0.12355911731719971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,128,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,768,0.0061048890153567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,768,0.00665155549844106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,3072,0.011472889118724398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,768,0.022662222385406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,16384,1.727269384596083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,512,0.5334515571594238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,4096,0.008984000318580205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,3584,0.6069546805487739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,2560,0.01815200017558204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,64,0.020388444264729817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,65536,0.20733333958519828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,12288,0.39073067241244847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,2048,0.01091822236776352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,4096,0.014621333943472968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,16384,0.05965422259436714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,10240,0.03808533483081394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,6144,0.10850044753816392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,10240,0.5321591165330675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,2560,0.010623999767833285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,5120,0.017163554827372234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,4096,0.19671200381384957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,5120,0.023892444041040208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,1536,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,2560,0.015751111838552687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,65536,0.48929066128200954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,2560,0.006351111249791251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,8192,0.015167999598715039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,2560,1.0220532947116427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,128,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,4096,0.007532444265153673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,8192,0.14186488257514104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,16384,0.054692443874147206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,3584,0.015467555986510383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,2048,0.011746666497654386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,10240,0.8318533367580838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,4096,0.09790133105383979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,3584,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,2560,0.010306666294733683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,4096,0.018265777164035372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,4096,0.21786133448282877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,6144,0.056001775794559054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,51200,0.19984711541069877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,1536,0.01034044474363327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,2048,10.497105068630642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,51200,0.12469955285390218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,16384,0.822682645585802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,512,0.00814666681819492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,5120,0.02660977840423584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,512,0.026928888426886663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,3072,0.009572444690598382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,64,0.012448000411192576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,5120,0.24914754761589897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,128,0.0042568888101312844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,512,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,8192,0.13003733423021105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,2048,0.0052800000541739995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,16384,0.44196444087558323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,10240,0.26210400793287486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,3072,0.00998044427898195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,7168,0.17799554930792916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,3584,0.007267555428875818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,16384,0.0743288861380683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,7168,0.12916444407569036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,65536,0.010665777656767102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,512,0.010684444672531553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,16384,0.9271102481418185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,128,0.006243555496136348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,64,0.03151822090148926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,256,0.013873777455753751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,16384,0.3269084559546577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,51200,0.19998044437832305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,2048,0.05103288756476509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,512,0.03358755509058634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,64,0.006687110910813014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,4096,0.08612355258729722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,12288,6.775881025526259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,6144,0.12743555174933538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,128,0.017251556118329365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,128,0.004283555679851108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,128,0.0047155556579430895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,1536,0.007183111376232571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,10240,0.03712177938885159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,256,0.020057777563730877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,16384,0.023019555542204116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,3584,0.05998311440149943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,5120,0.1673520008722941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,10240,0.09834221998850505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,7168,0.5135422282748753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,768,0.02168888847033183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,32,0.015068444940778943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,32,0.01071733319097095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,12288,0.01860000027550591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,8192,0.1297573381000095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,7168,0.009408888717492422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,4096,0.13876622252994114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,32,0.0026888888743188647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,256,0.007845333053006066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,2560,0.03774666786193848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,51200,0.13788533210754395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,2560,1.1291297276814778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,4096,0.10146222511927287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,7168,0.02741777731312646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,12288,0.09337777561611599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,512,0.02367377777894338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,3072,0.17784533235761854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,3072,0.03650044401486715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,12288,0.5799813270568848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,3072,0.017771555317772757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,32,0.030341333813137476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,5120,0.02104355560408698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,128,0.014813333749771118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,12288,6.252300262451172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,12288,0.02051733268631829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,51200,0.12358933024936253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,6144,0.14165600140889487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,5120,0.011938666303952536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,12288,0.506773312886556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,51200,1.5284177992078993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,1024,0.7415839831034342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,512,0.015356444650226168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,256,0.004087111188305749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,128,0.011169777976142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,4096,0.013231111069520315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,1024,0.053542223241594106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,3584,0.21203911304473877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,256,0.11754666434393989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,768,0.004351110921965705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,2048,0.009848000274764167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,7168,0.0661804477373759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,256,0.017606221967273288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,256,0.0047022220161226065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,51200,35.04961140950521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,5120,0.053764445914162524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,5120,0.17897066805097792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,2560,0.009483555952707926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,65536,0.2393724388546414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,2048,0.016093333562215168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,8192,0.011748444702890186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,768,0.005702222386995952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,5120,0.0893484420246548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,8192,0.05429155627886454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,6144,0.2097733285692003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,768,0.00598933340774642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,5120,0.17448622650570342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,3072,0.04122844338417053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,6144,0.16102933883666992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,2048,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,12288,0.08301244179407756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,256,0.0028311111446883944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,8192,0.15259377161661783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,7168,0.21445955170525444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,768,0.019595555133289762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,64,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,2048,0.007191999918884701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,32,0.004471111214823193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,10240,0.2273013326856825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,8192,0.026643556025293138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,16384,0.03867022196451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,1536,0.009026666482289633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,51200,0.7052551375495063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,7168,0.028805332051383123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,128,0.003543111185232798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,64,0.0026480000880029467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,2560,0.017090666625234816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,7168,0.045440001620186694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,16384,0.3285733328925239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,64,0.0025964443468385274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,7168,0.017859554953045316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,1536,0.013624889155228933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,3584,0.007832888927724626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,4096,0.02215555475817786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,7168,0.0064586665895250105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,7168,0.01090755561987559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,5120,0.014352889524565803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,2560,0.008032889001899296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,768,0.0038319999972979226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,64,0.005044444567627377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,3072,0.014365333649847241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,128,0.033891555335786604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,7168,0.02251822253068288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,51200,2.3399920993381076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,32,0.006711111300521427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,8192,0.3727253278096517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,768,0.027752000424597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,8192,0.7337199846903483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,2048,0.007458666960398356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,16384,0.3177013397216797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,5120,0.007720000214046902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,1024,0.00701422244310379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,5120,0.06068711148367988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,8192,0.016149333781666227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,1024,0.011954666839705573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,256,0.003216000066863166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,256,0.005166222237878376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,6144,0.11922755506303574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,256,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,2048,0.016941333810488384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,65536,0.5213768747117784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,2560,0.022864888111750286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,128,0.005193777796294955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,32,0.0034151110384199354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,65536,0.08654489119847615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,3584,0.00996088898844189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,2048,0.017255110873116385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,128,0.008589333130253686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,51200,0.01680799987581041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,4096,0.09621333413653904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,65536,0.06914222240447998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,4096,0.46260976791381836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,3072,0.2919955518510607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,3072,0.012524444195959302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,12288,0.15206489298078749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,5120,0.13046488496992323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,1024,0.006911110960774952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,10240,0.2843342092302111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,768,0.005019555489222209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,768,0.0041351111398802865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,16384,0.22947377628750273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,128,0.12187643845876057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,5120,0.009366222553782994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,5120,0.06740711132685344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,7168,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,16384,0.022582223018010456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,128,0.00544799988468488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,512,0.266774230533176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,128,0.024513777759340074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,10240,0.014551111393504672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,256,0.015809777710172866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,10240,0.016608888904253643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,4096,7.1324708726671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,32,0.0032915555768542816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,256,0.005819555371999741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,512,0.009160888691743216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,3072,0.11450133058759902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,3072,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,10240,0.43204087681240505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,32,0.006572444405820634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,5120,0.07986933655209012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,1536,0.03238044513596429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,2560,0.012977778083748288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,4096,0.17241333590613472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,7168,3.4393865797254772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,768,0.08900266885757446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,3072,0.5287537574768066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,1536,0.011564444336626263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,768,0.09672177500194973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,5120,0.05480800072352091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,5120,0.3002213372124566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,512,0.04048444496260749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,128,0.1587217781278822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,1024,0.028528001573350694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,4096,0.01130311108297772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,16384,0.008400888906584846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,2560,0.014919999572965833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,32,0.004447111239035924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,256,0.004483555754025777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,8192,0.3255777888827854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,2560,0.00620266670982043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,10240,0.018264888061417472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,12288,0.21364443831973604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,8192,0.21210843986935088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,10240,0.015544889701737298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,256,0.033166223102145724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,2560,0.13862755563524035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,16384,0.13824266857571071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,2560,0.011879111329714457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,65536,0.26758400599161786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,51200,6.145045386420356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,3072,0.028130667077170476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,1536,0.01313066648112403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,64,0.01947022146648831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,256,0.165157331360711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,32,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,1024,0.012387555506494312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,6144,0.03308711118168301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,512,0.011932444241311816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,1024,0.037696888049443565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,3072,0.019820445113711886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,512,0.11765421761406793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,4096,0.11220888296763103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,8192,0.018990221950742934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,4096,0.013376889129479727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,256,0.0070186663005087115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,16384,0.022036444809701707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,8192,0.011004444625642566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,7168,0.010580444501505958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,1536,0.006215111249023014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,2048,0.02959644463327196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,10240,0.019742222295867074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,1536,0.007243555453088548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,64,0.002711999954448806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,32,0.00526311124364535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,51200,0.07225155830383301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,2048,0.039656887451807656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,256,0.020266667008399963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,8192,0.08632444673114353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,768,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,1024,0.025237333443429735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,4096,0.010399999717871347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,7168,0.011153777440388998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,512,0.0032960000551409195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,5120,0.008716444174448649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,2560,0.020391111572583515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,2560,0.08465066883299087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,65536,0.2693866623772515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,7168,0.0358240008354187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,2048,0.040258665879567467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,3072,0.14981244670020208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,512,0.016698666744761996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,3584,0.05758844481574165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,128,0.015252444479200574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,65536,0.9318781958685981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,2048,0.17351110776265463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,16384,0.035980443159739174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,16384,0.04932711190647549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,3072,0.008815999660227034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,6144,0.3393093215094672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,8192,0.03776266508632236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,64,0.003399111123548614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,3584,0.07887022362815009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,5120,0.016647110382715862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,2048,0.016844444804721408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,7168,0.08439199792014228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,7168,0.009443555441167619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,1536,0.01739644507567088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,768,0.011962666279739805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,2560,0.14550933572981092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,128,0.006899555524190267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,512,0.0041271112859249115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,3072,0.011337777806652917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,128,0.004293333325121137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,51200,0.9673902723524305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,768,0.015200000670221118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,2048,0.01583377851380242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,3072,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,16384,0.6940595308939616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,7168,0.0587057802412245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,128,0.017047999633683097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,2560,0.010716444088353051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,2048,0.016511110795868766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,512,0.007824000385072496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,3072,0.01015466699997584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,65536,0.29381243387858075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,3584,0.33811288409762913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,1024,0.5831795798407661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,4096,0.04080355498525832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,32,0.0027546667390399506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,512,0.0038657776183552216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,512,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,5120,0.15961421860588923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,512,0.003573333223660787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,256,0.004322666674852371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,512,0.0037831109431054858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,768,0.02998044424586826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,3584,0.12077066633436416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,512,0.00777599960565567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,3072,0.0865866674317254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,7168,0.05660711394415962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,10240,0.013592888911565145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,128,0.006166222194830577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,12288,0.021320889393488567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,128,0.0032400000426504347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,2048,0.1092453334066603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,12288,0.020776000287797716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,4096,0.25716710090637207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,3584,1.1200133429633248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,51200,0.05714933077494303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,8192,0.01387288918097814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,1024,0.0058328890138202244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,2048,0.07347022162543403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,4096,0.01917155583699544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,3584,0.01853511068556044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,7168,0.8001662360297309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,16384,0.03252089023590088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,3584,0.017032000753614638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,2048,0.007607999775144789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,7168,0.39676178826226133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,256,0.009087111386987898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,3584,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,512,0.015649777319696214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,2048,0.005349333501524395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,3584,0.15282489193810359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,51200,4.126895056830512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,512,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,1024,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,65536,0.9030115339491102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,1024,0.010956444674068026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,1536,0.007440889047251807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,2560,0.010847999817795224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,65536,28.82833523220486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,16384,0.055418666866090566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,1024,0.013368888861603208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,65536,0.12350577778286403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,512,0.03122044603029887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,3584,0.2386239899529351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,7168,0.10547555817498101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,65536,2.6366871727837458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,8192,0.016535111599498324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,5120,0.00959822204377916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,256,0.033999111917283796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,2048,0.015047111445003085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,3072,0.013289777768982781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,2560,0.24063377910190156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,7168,0.014888889259762235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,128,0.004980444494220945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,10240,0.018183110488785636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,3072,0.008187555604510838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,512,0.005639111002286275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,8192,0.19833955499860975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,32,0.007638221813572778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,3584,0.009834667046864828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,4096,0.022014222211307947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,16384,0.5789013438754612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,128,0.005869333528810077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,5120,0.23024445109897188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,12288,2.721201790703667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,2048,0.04211022125350105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,512,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,64,0.0030586665703190696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,256,0.009870222045315636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,1536,0.005701333284378052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,65536,1.3556533389621312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,10240,0.08034488889906141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,2560,0.057414220439063184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,5120,0.03431111243036058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,2560,0.01276711126168569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,7168,0.06400799751281738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,6144,0.5219457944234213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,1024,0.23267110188802084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,128,0.004353777815898259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,8192,0.09581422143512303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,64,0.002936888900068071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,5120,0.014833778142929077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,768,0.0037813331517908308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,51200,0.04403822289572822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,8192,0.027170666389995154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,256,0.005470222069157495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,5120,0.024304888314670984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,1536,0.006693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,128,0.009159111314349705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,16384,0.2694844404856364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,1536,0.05470755696296692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,65536,0.14851555559370253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,512,0.023111111587948267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,64,0.004813333352406819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,3584,0.019135110908084445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,1024,0.007143110864692264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,1024,0.004351110921965705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,12288,0.5704293251037598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,256,0.009141333401203156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,3584,1.3652612898084853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,12288,0.39592178662618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,768,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,16384,1.0224808586968315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,1536,0.5484728813171387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,16384,0.025431111454963684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,12288,0.03882577684190538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,6144,0.017298666967286002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,51200,0.01278222186697854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,4096,0.00889155517021815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,8192,0.045258667733934194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,8192,2.2605813344319663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,7168,0.028792887926101685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,512,0.0058613332609335584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,1024,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,2048,0.007958222594526079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,4096,0.2540035512712267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,2048,0.007320889168315464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,6144,0.04197777642144097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,8192,0.023657777243190344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,256,0.006660444455014334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,256,0.011169777976142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,5120,0.01868711080816057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,65536,0.14759555127885607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,65536,0.9237448374430338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,12288,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,1024,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,65536,0.6916497548421224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,2560,0.04370933439996508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,1536,0.16769866148630777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,3072,0.09396088785595363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,32,0.008997333546479544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,51200,0.22412088182237414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,128,0.0052986666560173035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,4096,0.17423910564846465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,768,0.003998222036494149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,16384,0.006370666540331311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,32,0.016006223029560514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,5120,0.05843200286229452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,3072,0.13660177919599745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,10240,0.01555288831392924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,7168,0.1251937813229031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,3072,0.04712266723314921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,256,0.005421333428886201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,65536,0.06305333640840319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,3584,0.06251733170615302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,7168,0.017471111483044095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,6144,0.02404800057411194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,10240,3.594841851128472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,1024,0.005946666830115848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,128,0.003683555457327101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,128,0.0037271111375755737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,32,0.011047110789351992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,3584,0.012011555333932241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,512,0.005971555494599872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,5120,0.18093332979414198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,3584,0.009555555880069733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,51200,4.722676595052083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,5120,0.018315555320845712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,10240,0.1643857823477851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,128,0.047707554366853505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,12288,0.6061600049336752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,12288,0.049513777097066246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,3584,0.009196444518036312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,3584,0.033801777495278254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,6144,0.09236444367302789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,32,0.01273155543539259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,6144,0.09510933028327094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,6144,0.1965564489364624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,8192,2.428356382581923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,1536,0.15222133530510798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,128,0.005070222334729301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,12288,0.22761066754659018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,6144,0.020219556159443326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,8192,0.017575111654069688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,5120,0.4197564389970567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,4096,0.07409244113498263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,8192,0.12420354949103461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,1536,0.05131999982727898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,128,0.0072266666425599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,2560,0.014136888914638095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,51200,21.081639607747395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,2560,0.09285244676801895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,768,0.03958488835228814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,64,0.007638221813572778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,64,0.002944888960984018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,65536,0.05495466788609823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,5120,0.02787733409139845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,12288,0.06186044216156006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,2048,0.20484266016218397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,768,0.8008284568786621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,512,0.006923555499977536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,16384,0.05344444513320923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,16384,0.09740177790323894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,2560,0.018753778603341844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,51200,0.28204088740878636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,1024,0.03868800070550706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,4096,0.6074853473239475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,12288,0.10471644666459824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,256,0.011727111207114326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,4096,0.0373288889726003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,512,0.03822755482461717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,8192,0.11042577690548366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,5120,0.007620444728268518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,128,0.008267555799749162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,64,0.006526222245560751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,8192,0.022133333815468684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,10240,0.015014222926563688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,3584,0.2140995528962877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,2048,0.010430222584141625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,10240,0.022218666142887537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,512,0.006648889018429651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,64,0.003481777591837777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,5120,0.1338933308919271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,12288,0.05740177631378174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,8192,0.05019022358788384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,2560,0.01615111033121745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,256,0.029316445191701252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,5120,0.39203821288214785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,7168,0.4200942251417372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,64,0.006461333483457565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,3072,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,512,0.007255111303594377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,256,0.003108444520168834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,3584,0.009956444303194681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,4096,0.018541332748201158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,16384,0.2155359983444214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,768,0.01573866605758667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,768,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,16384,0.03432888785998026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,768,0.01351199961370892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,8192,0.012087999946541257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,65536,0.17590399583180746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,2048,0.06817599799897936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,8192,0.014104000396198697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,5120,0.006400888992680444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,8192,0.03302222159173753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,2048,0.009371555513805812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,256,0.003042666655447748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,4096,0.02144355575243632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,16384,0.021945777866575453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,5120,0.011576889289749993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,1536,0.008199111455016667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,8192,0.11367021666632758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,7168,0.2225031057993571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,6144,0.026123555170165166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,10240,0.24018756548563638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,5120,0.09658666451772054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,7168,0.01369244439734353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,768,0.008512000242869059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,128,0.006212444355090459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,65536,0.00904444439543618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,2560,0.007689778175618913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,768,0.01535999940501319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,2048,0.017503110898865592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,8192,0.0240631103515625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,4096,0.01942044496536255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,3584,0.028661333852344092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,1024,0.004078222231732475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,12288,0.13070755534701878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,3584,0.012056888805495368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,2560,0.007041778001520369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,5120,0.013448889056841532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,16384,0.23737510045369467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,2048,0.007369777394665613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,65536,1.103282716539171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,6144,0.16654755009545222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,4096,0.01740533279048072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,768,0.004690666579537922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,65536,0.06688888867696126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,512,0.00554044461912579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,64,0.007920888562997183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,65536,0.2503395610385471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,6144,0.018775999546051025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,2560,0.035565333233939275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,5120,0.024463999602529738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,512,0.005384888913896348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,4096,0.011936888926559024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,2048,0.006648889018429651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,128,0.02935911218325297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,12288,1.0765537685818143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,512,0.047673778401480786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,10240,0.016843555702103507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,512,0.007164444360468123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,32,0.010082667072614035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,51200,0.1118968857659234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,512,0.005168888717889786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,10240,0.03179288903872172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,1024,0.03904711206754049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,7168,0.13846577538384333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,6144,0.06991200314627753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,1024,0.005097777893145879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,128,0.002814222127199173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,3072,0.02188266648186578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,1536,0.004522666749027041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,2560,0.05682933330535889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,10240,0.48482486936781144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,6144,0.020106666617923312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,128,0.008100444244013892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,51200,25.02780320909288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,65536,2.055013232760959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,12288,0.06789066394170125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,12288,1.1170132954915364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,4096,0.4848151206970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,2048,0.1480613284640842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,12288,0.02215733296341366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,2560,0.06798933611975776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,32,0.005561777700980504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,5120,0.018684445155991446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,32,0.0034248888906505373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,3584,0.019891555110613506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,10240,1.4808870951334636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,51200,0.18622844749026826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,2560,0.02440711193614536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,1024,4.027478959825304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,3072,0.06736089123619927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,3584,0.013447999954223633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,768,0.007320000065697565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,2048,0.031164444155163232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,5120,0.11976710955301921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,256,0.003301333429084884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,2048,0.04862933357556661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,64,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,512,0.018047110901938546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,51200,2.8250880771213107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,51200,0.29491999414232045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,8192,0.39461954434712726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,2560,0.03247111042340597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,1536,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,4096,0.011744000017642975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,128,0.002871111035346985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,4096,0.008031110796663497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,512,0.006311111152172089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,2560,1.2919075224134657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,6144,0.07515377468532987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,16384,0.07760710848702325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,4096,0.008438222110271454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,768,0.019868445065286424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,1536,0.005199999858935674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,5120,0.010791999598344168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,16384,0.029993779129452173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,1536,0.07285688983069526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,51200,0.09800089067882961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,768,0.0036862223512596558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,2048,0.010434666441546546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,128,0.10349422030978733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,3072,0.008774222599135505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,3584,0.012351999680201212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,4096,0.10006578101052178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,5120,0.02330933345688714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,51200,1.6841066148546007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,65536,2.555121739705404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,4096,0.025411556164423626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,64,0.00817599975400501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,16384,0.09187910954157512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,6144,0.01014755583471722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,2560,0.044418666097852916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,64,0.04811199837260776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,2560,0.007696000238259633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,12288,0.021225778592957392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,3072,0.02775466607676612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,6144,0.3049315611521403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,512,0.007182222273614671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,51200,0.009882666998439366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,7168,0.01744622157679664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,3072,0.024012444747818842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,512,0.039705776506000094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,65536,0.30969421068827313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,2560,0.014169777433077494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,51200,15.190241495768229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,12288,0.024671110841963027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,51200,0.14324355125427246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,12288,0.5250444412231445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,3584,0.014339556296666464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,7168,0.11068799760606553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,5120,0.01551466683546702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,2560,0.011042666931947073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,12288,0.4326959980858697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,2048,0.05827733543184069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,2048,0.01719377769364251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,64,0.0030337776988744736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,32,0.008229333493444655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,3072,0.03162399927775065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,5120,0.04262222184075249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,8192,0.047875556680891246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,2048,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,768,0.024334222078323364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,128,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,2560,0.01919911139541202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,3584,0.012487110992272695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,12288,7.4064441257052955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,5120,0.9139706293741862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,10240,0.03747555613517761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,2560,0.008037333687146505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,4096,0.012320888539155325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,6144,0.051501333713531494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,51200,0.5371297730339898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,2048,0.0075128889746136135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,768,0.3742826779683431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,2048,0.06015111340416802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,6144,0.019187556372748483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,4096,0.016332444217469957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,512,0.005944889038801193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,64,0.0026568888376156488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,128,0.005225777626037598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,3584,0.2577093442281087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,1024,0.028705779049131606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,4096,0.019377777973810833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,6144,0.04163466559516059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,2560,0.018350222044520907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,768,0.006292444550328785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,8192,0.012109333442317115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,768,0.07702044645945232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,16384,0.024187554915746052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,6144,0.021291555629836187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,3584,0.01953066719902886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,1024,0.03592888845337762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,16384,0.5329457918802897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,128,0.0034026667061779234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,64,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,768,0.00481599983241823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,12288,21.911173502604168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,16384,6.348423851860894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,1536,0.007659555309348636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,256,0.010234666367371878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,8192,0.13565511173672146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,2560,0.011010666688283285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,768,0.005110222018427319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,6144,0.1599982182184855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,8192,0.3615208996666803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,3072,0.012109333442317115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,1536,0.05473688907093472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,1024,0.00637155564294921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,10240,0.026160889201694067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,3584,0.05040355523427328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,256,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,8192,0.01148444414138794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,2560,0.07503200239605375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,65536,0.025288888149791296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,512,0.2681982252332899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,1536,0.028813332319259644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,4096,0.02720444401105245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,16384,0.11009155379401313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,512,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,512,0.008488888541857401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,768,0.02981511089536879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,12288,0.022279111875428095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,768,0.30585066477457684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,32,0.0027351110345787476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,5120,0.009202666580677032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,65536,0.23479554388258192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,10240,0.017687110437287223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,7168,0.04117066661516825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,512,0.0032924444725116095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,8192,0.07463022073109944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,7168,0.27962756156921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,3584,0.009106666677527958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,2560,0.01200977795653873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,1536,0.006294222341643439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,256,0.00518222235971027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,10240,0.05742400222354465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,64,0.005098666581842634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,12288,0.028318223026063707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,128,0.008730666504965888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,65536,0.29803289307488334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,1536,0.01791466606987847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,256,0.009714666340086196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,1024,0.012324444121784635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,256,0.003116444374124209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,5120,0.050719112157821655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,6144,0.026679999298519556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,2560,0.08081422249476115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,3584,0.01220444424284829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,6144,0.05189244283570183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,65536,6.948543972439236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,65536,0.1874799993303087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,768,0.022152000003390845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,6144,0.02482755482196808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,12288,0.0324942237801022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,2560,0.35808799001905656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,3072,0.015899555550681222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,51200,91.04759385850694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,51200,3.264906565348307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,3584,0.0745226674609714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,12288,0.02438666588730282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,4096,0.0081395556529363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,2560,0.20591021908654106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,1536,0.008960888617568547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,8192,0.02494666642612881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,3584,0.014898666077189975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,10240,7.436427646213108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,512,0.005036444299750858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,1536,0.005600000007285013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,768,0.012063110868136087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,2560,0.010391111175219217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,1536,0.016759999924235873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,768,0.03779733180999756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,1024,0.00647288892004225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,3584,0.013570666313171387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,1024,0.008652444514963362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,32,0.005694222119119432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,6144,0.19269688924153647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,64,0.0032026666320032547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,10240,0.3393084473080105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,32,0.0029333333174387612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,2048,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,4096,0.02223288847340478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,65536,0.08849422136942546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,65536,1.076175159878201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,1024,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,51200,21.17958747016059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,128,0.004857777721352047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,256,0.0038497779104444715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,10240,0.03325244453218248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,6144,0.021531555387708876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,3584,0.029102222787009344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,1536,0.026527111728986103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,5120,0.019191111127535503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,2560,0.03910488883654276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,6144,4.430675506591797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,2560,0.02407377792729272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,3072,0.13443556096818712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,8192,0.026219555073314246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,256,0.014242667290899487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,1536,0.13929333951738146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,1536,0.010026666853162978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,3072,0.1059875554508633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,256,0.0063795554969045855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,1024,0.005383111122581694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,2560,0.00980000032318963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,3584,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,7168,0.5803457895914713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,12288,0.1924826701482137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,16384,0.03998400105370416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,51200,0.0481644438372718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,16384,0.01942933268017239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,16384,0.23603643311394584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,2560,0.013206221991115145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,1024,0.00488800017370118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,2048,0.018453333112928603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,4096,0.0843048890431722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,65536,2.190763473510742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,12288,0.020240889655219186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,3584,0.014933332800865173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,1024,0.003990222182538775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,256,0.007865777446164025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,12288,0.44941065046522355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,65536,0.401142226325141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,512,0.010846222440401712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,1536,0.007801777786678738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,5120,0.07558221949471368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,51200,0.23591200510660806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,16384,0.022641777992248535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,256,0.02870933214823405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,12288,0.7588897811041938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,128,0.015563555889659457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,8192,0.02819288770357768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,8192,0.3528711001078288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,2560,0.008108444511890411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,16384,0.07364977730645074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,2048,0.011767110890812345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,256,0.02416444487041897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,65536,0.17443377441830107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,4096,0.02737688852681054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,7168,0.3005795478820801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,1024,0.029751112063725788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,2048,0.012514666550689273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,2048,0.030100444952646892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,4096,0.009306666751702627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,3584,0.07029600275887384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,2560,0.007763555480374231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,4096,0.012356444365448423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,1024,0.003751111113362842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,6144,0.178384886847602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,128,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,2560,0.01807466646035512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,4096,0.008373333348168267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,5120,0.016757332616382174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,7168,0.04537689023547702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,10240,0.24831822183397081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,3584,0.3049822118547228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,128,0.00319377767542998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,1024,0.01520088811715444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,256,0.008008889026112026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,768,0.005333333379692501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,64,0.002672000063790215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,16384,0.8984240425957574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,256,0.005853333406978183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,7168,0.025676444172859192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,768,0.004278222305907143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,3072,0.05421688821580675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,32,0.0125946667459276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,6144,0.1180071168475681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,2048,0.014300444059901767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,768,0.011262222296661802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,8192,0.26331112119886607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,6144,2.984546661376953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,1536,0.0064995553758409284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,5120,0.02730133301681942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,5120,0.03151199883884854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,16384,0.36434221267700195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,32,0.030504888958401147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,3072,0.03092355529467265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,2560,0.13657332791222468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,2048,0.010394666757848527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,12288,0.017450667089886136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,768,0.004077333129114575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,5120,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,4096,0.03495466709136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,64,0.007685333490371704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,2560,0.01184622198343277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,7168,0.025332444243960913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,768,0.00666133314371109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,12288,3.6589342753092446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,1024,0.009786666267448002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,256,0.00674933319290479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,128,0.005154666801293691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,3072,0.014290667242474027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,32,0.01295377810796102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,256,0.005675555517276128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,3584,0.01460355520248413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,12288,0.021736888421906367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,1024,0.00573333352804184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,8192,0.03284444411595663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,32,0.003959111041492886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,12288,0.050155556864208646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,5120,0.010096000300513374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,8192,0.29594577683342826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,512,0.0033359999457995095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,10240,4.457464853922526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,7168,0.018815999229749043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,256,0.005160000175237656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,12288,0.18079200055864122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,12288,0.21534844239552817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,16384,7.493565029568142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,256,0.03682400120629205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,65536,8.47055901421441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,2560,0.009199110998047723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,32,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,32,0.011138666835096149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,2560,0.04668088754018148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,12288,0.08917422427071466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,4096,0.1838231086730957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,768,0.033914665381113686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,8192,0.0726791090435452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,768,0.00590133335855272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,51200,0.013789333403110504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,256,0.010245333115259806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,8192,0.045384890503353544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,51200,0.7575768894619412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,51200,0.06952799691094293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,6144,0.9652960035536023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,65536,0.5650773578219944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,65536,0.35820889472961426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,512,0.010460444622569613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,10240,2.8225173950195312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,3072,0.017027555240525138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,16384,0.024791111548741657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,1536,0.007063999772071838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,256,0.0063288890653186375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,8192,21.742351955837673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,1024,0.04524088899294535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,1536,0.007865777446164025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,64,0.007035555938879649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,5120,0.22787822617424858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,3072,0.12794755564795599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,5120,0.484208001030816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,3584,0.011855111353927188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,8192,0.015189333094490899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,3072,0.23871289359198677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,128,0.0158942225906584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,10240,0.016750221451123554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,512,0.005889777921968036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,51200,0.041703111595577665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,64,0.023266666465335425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,12288,2.3281688690185547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,7168,0.32230933507283527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,10240,1.173842642042372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,32,0.00868711123863856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,128,0.007051555646790399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,12288,0.4016755686865912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,3072,0.8620497915479871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,8192,0.3955493238237169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,51200,0.06435022089216444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,768,0.014196443888876172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,7168,0.00612088872326745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,1024,0.007041778001520369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,2048,0.024029332730505202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,10240,0.2688026693132189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,1024,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,12288,0.03811555438571506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,2560,0.1029591096772088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,5120,0.020968000094095867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,5120,0.020717332760492962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,8192,0.017027555240525138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,10240,0.038527111212412514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,64,0.007073777417341868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,10240,0.04640088809861077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,5120,0.012171555724408893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,2560,0.00832622249921163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,5120,0.013016888664828407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,5120,0.06454666455586751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,65536,0.06931555271148682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,1536,0.009033777647548253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,6144,0.13011466132269964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,1024,0.07434755563735962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,128,0.0027786667148272195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,768,0.005674666828579373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,3584,0.22294312053256563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,65536,1.8058355119493272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,768,0.01613066593805949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,64,0.006053333481152852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,4096,0.02343911098109351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,128,0.2879893249935574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,1536,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,256,0.0031191110610961914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,3072,0.0699288911289639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,64,0.0030177777840031516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,1024,0.006015111174848344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,3584,0.03812711106406318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,65536,0.2720115449693468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,64,0.002612444468670421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,51200,0.1306862168841892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,8192,0.01626755628320906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,4096,0.14098221725887722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,3584,0.008650666309727563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,51200,1.900345696343316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,3584,0.2278613249460856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,3072,0.022904000348514978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,10240,0.34592800670199925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,51200,0.015319110618697273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,10240,0.04972622129652235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,8192,0.017456000049908955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,768,0.07314311133490668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,16384,0.03750844465361701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,128,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,10240,0.033223999871148005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,65536,0.22983998722500273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,5120,0.02109955582353804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,512,0.004575110971927643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,4096,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,4096,0.08647644519805908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,7168,0.020469332734743755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,8192,0.014228444960382251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,4096,0.03140799866782294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,8192,0.04828977915975782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,51200,0.061298668384552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,6144,0.08293777704238892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,5120,0.023939554889996845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,64,0.005346666607591841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,512,0.008564444051848518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,5120,0.008948444492287105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,6144,0.12234755357106526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,12288,0.087391111585829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,16384,1.2509902318318684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,3584,0.04976000057326423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,12288,0.02190133266978794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,128,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,7168,0.10856088664796616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,32,0.0052986666560173035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,2560,0.006078222145636876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,32,0.011970666547616323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,3584,0.027649778458807204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,1536,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,768,0.005746666755941179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,65536,0.48617156346638996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,65536,0.1350488927629259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,10240,0.029461334149042766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,3584,0.03492355677816603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,256,0.0031582222630580268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,3584,0.010741333166758219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,3584,0.01330311099688212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,51200,0.053674668073654175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,6144,0.008935111264387766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,3072,0.012702222499582501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,2048,0.0628933310508728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,256,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,3072,0.006404444575309753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,256,0.007658667034573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,64,0.005113777601056629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,3584,0.010000000397364298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,768,0.008991999758614434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,2560,0.09142933289210002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,2560,0.05065422256787618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,6144,0.0920320020781623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,8192,0.06556799676683214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,1024,0.005594666633341048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,256,0.004508444418509801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,51200,0.03671733207172818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,3072,0.26451465818617076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,256,0.006275555739800136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,5120,0.009513777991135916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,128,0.004693333473470476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,2048,0.014378666877746582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,51200,2.050455093383789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,65536,0.17876622411939833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,1536,0.014811555544535318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,7168,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,5120,0.023043556345833674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,51200,0.12481155660417344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,12288,0.03540622194608053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,12288,0.27549510531955296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,51200,3.5962914360894094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,16384,0.350508451461792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,51200,1.729116439819336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,512,0.01690933273898231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,3072,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,10240,0.014877332581414117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,256,0.009008000294367472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,10240,0.27342134051852757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,2048,0.037008888191647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,128,0.003779555360476176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,3584,0.047919111119376294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,10240,0.03336888882848952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,3584,0.0090524446633127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,2048,0.008456889126035903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,4096,0.023883556326230366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,128,0.005084444251325396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,51200,19.616972181532116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,3584,0.014147554834683737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,7168,0.1340915494494968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,64,0.006047111004590988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,1536,0.10976711246702407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,128,0.010787555740939246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,1536,0.36100443204243976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,16384,0.06205422348446316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,12288,0.08274488978915744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,128,0.007873777714040544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,16384,0.10510755247539944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,4096,0.007658667034573025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,10240,0.013771555489963956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,51200,0.21560000048743355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,3584,0.029213332467608984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,12288,0.10362666845321655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,1536,0.0069439998931354946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,32,0.006558222075303395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,3072,0.07531555493672688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,12288,0.03249955508444045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,32,0.005576888720194499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,65536,0.2627120018005371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,768,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,64,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,8192,0.019705777366956074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,16384,0.02158755560715993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,10240,0.01037688884470198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,65536,0.2561466693878174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,7168,0.014195554786258273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,3584,0.07059644328223334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,3072,0.02271555529700385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,12288,0.02494755552874671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,768,0.004776000148720211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,10240,0.017859554953045316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,2560,0.0074515557951397365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,32,0.005164444446563721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,10240,3.0599831475151906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,128,0.004159111115667555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,10240,4.934050665961371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,4096,0.01239377756913503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,8192,0.014190221826235453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,4096,0.013357333011097379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,3584,0.15351111359066433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,5120,0.14952622519599065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,64,0.02915466825167338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,12288,0.5296808878580729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,10240,0.271579556994968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,1024,0.006486222147941589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,1024,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,256,0.015045333239767285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,16384,0.17093066374460855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,64,0.0030248889492617715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,51200,0.18317600091298422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,256,0.00829777783817715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,3584,0.21152710914611816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,5120,0.12021866109636094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,128,0.20533955097198486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,512,0.007126222054163615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,32,0.003999111139112049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,64,0.004289777742491828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,3584,0.010273777776294285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,5120,0.00925688859489229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,65536,0.31109688017103404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,16384,0.14002399974399143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,51200,3.8404354519314237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,4096,0.016880000631014507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,3584,0.04056444433000352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,768,0.025039111574490864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,51200,0.014550222290886773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,3072,0.015263110399246216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,16384,0.10040533542633057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,5120,1.3711004257202148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,64,0.006793777975771163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,64,0.045253333118226796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,1024,0.006542222367392645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,128,0.010847110715177325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,3584,0.014801777071423002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,51200,0.16992533206939697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,512,0.009635555247465769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,1536,0.006911110960774952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,65536,0.4882399770948622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,65536,0.26236621538798016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,256,0.01573511130279965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,7168,0.006015111174848344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,2560,0.11020977629555596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,1024,0.05829155445098877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,128,0.002708444371819496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,16384,0.18452178107367623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,1536,0.02908710969818963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,16384,0.2820097870296902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,51200,0.562711132897271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,10240,0.01725955473052131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,8192,0.03171022070778741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,1536,0.012293332980738746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,3584,0.006800000038411882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,2560,0.02734222180313534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,3072,0.009933333430025313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,256,0.042325334416495435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,3584,0.029283556673261855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,16384,0.46719201405843097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,2560,0.01593866613176134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,2048,0.016945777667893302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,2560,0.05144177873929342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,2048,0.13357510831620958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,256,0.016228444046444364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,2048,0.015511110424995422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,256,0.037492444117863975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,128,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,3072,0.009758222434255812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,64,0.005142222262091107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,8192,0.019932443896929424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,256,0.005120000077618493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,4096,0.10864088932673137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,6144,0.04057066639264425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,1536,0.006976889239417181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,6144,0.010024888647927178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,64,0.002549333290921317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,7168,2.5214977264404297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,1024,0.023416888382699754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,5120,0.012776888906955719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,1536,0.2465146647559272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,256,0.008415111237102086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,512,0.009747555686367882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,12288,0.1847928894890679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,6144,0.04279911186960009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,128,0.004495111190610462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,12288,0.09805778000089858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,8192,0.39798667695787215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,12288,0.3173777792188856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,2048,0.06909600231382582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,2560,0.12040800518459743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,65536,0.3547351095411513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,6144,0.07265600230958727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,5120,0.05711466736263699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,16384,0.035094221433003746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,32,0.00674933319290479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,512,0.00509333320789867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,768,0.012280888855457306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,7168,0.47031376096937394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,5120,0.018013333280881245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,5120,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,2048,0.013669333524174161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,3072,0.008351999852392409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,7168,0.07921155293782552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,5120,0.06931289037068684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,2048,0.012166221936543783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,4096,0.028928889168633357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,5120,0.01624000072479248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,1024,0.009290667043791877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,12288,0.12328533331553142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,8192,0.5789377954271104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,12288,0.04909244510862562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,512,0.0053484443989064955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,128,0.01331377774477005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,256,0.003121777748068174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,16384,0.04500088757938809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,6144,0.02122755514250861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,65536,0.066557334529029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,512,0.018910222583346896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,256,0.0507351126935747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,51200,0.08659288618299697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,256,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,2048,0.615091535780165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,64,0.04040711124738058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,2048,0.015967110792795818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,1536,0.006511999915043513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,3584,0.010267555713653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,2048,0.028513779242833454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,2560,0.008577778107590145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,768,0.02674755619631873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,3584,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,7168,0.04279466801219516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,6144,0.022805333137512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,10240,0.0395431121190389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,3072,0.047305779324637524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,2048,0.0123848890264829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,6144,0.006792888873153263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,2560,0.007308444215191736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,768,0.04685777756902906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,1024,0.03271911210483975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,8192,0.1194960011376275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,8192,0.16784889168209502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,2048,0.009161777794361115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,512,0.017900443739361234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,64,0.005586666779385672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,5120,0.09107110897699992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,1536,0.004877333425813251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,6144,0.017089777522616915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,51200,0.7673999998304578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,8192,0.014641778336630927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,8192,0.12434844175974528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,3072,0.007264888948864407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,10240,0.040539556079440646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,3072,0.1149653328789605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,65536,4.597677442762587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,512,0.004774222357405556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,768,0.007476444045702617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,128,0.006502222269773483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,7168,0.18198133839501274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,1536,0.00852888905339771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,4096,0.008416888614495596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,65536,0.3563724358876546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,10240,0.24510134590996635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,65536,0.1278302272160848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,16384,0.06933955351511638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,65536,4.175299750434028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,256,0.05453333258628845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,32,0.002495999965402815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,3072,0.007853333320882585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,3584,0.014721777704026965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,512,0.0047733332547876565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,64,0.0028026666906144884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,2048,0.034733331865734525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,3072,0.015288000305493673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,4096,0.25455111927456325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,65536,0.07231199741363525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,65536,0.029746668206320867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,10240,0.11865777439541286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,1536,0.006128888991143968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,6144,0.0184879998366038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,65536,0.0941386686431037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,2048,0.009502222140630087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,32,0.00546577779783143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,64,0.002761777697337998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,3072,0.10427555772993301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,512,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,1024,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,512,0.005623999983072281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,3584,0.008373333348168267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,5120,2.90940793355306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,128,0.022727999422285292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,3072,0.010787555740939246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,1536,0.07205955849753486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,128,0.006509333435032103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,3584,0.011477332976129321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,3072,0.032582223415374756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,6144,0.946565310160319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,512,0.008814222282833522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,128,0.005165333135260476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,3072,0.019714666737450492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,3072,0.11963466803232829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,5120,0.02986488739649455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,3072,0.03132177723778619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,1024,0.024930665890375774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,2048,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,2560,0.07805955410003662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,16384,0.021428444319301184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,256,0.00332266671790017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,6144,0.06411111354827881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,6144,7.404786851671006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,5120,0.047220445341534086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,1024,0.011350221931934357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,256,0.04616444309552511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,3584,0.014094221923086377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,51200,0.4513937632242839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,768,0.02290577855375078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,768,0.022816888160175745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,12288,0.09888444344202678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,4096,0.00830844458606508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,5120,0.014098667436175875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,2560,0.00870222184393141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,128,0.0031679999083280563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,3584,0.016533333394262526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,256,0.014212444424629211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,8192,0.007222221957312689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,10240,8.854932996961805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,51200,0.19876000616285536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,512,0.0042364444169733255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,8192,0.022307554880777996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,65536,0.5322728686862522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,64,0.005613333235184352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,3584,0.06834399700164795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,3584,0.008512000242869059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,3072,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,128,0.0048177776237328844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,12288,0.6165982352362739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,10240,0.649653328789605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,7168,0.2942897743648953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,7168,0.0321342216597663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,2560,0.06286488638983832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,256,0.19475466675228545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,6144,0.010254222485754225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,3584,0.007346666521496243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,51200,0.17010488775041369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,6144,0.017634666628307767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,51200,0.39536889394124347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,16384,0.0387715564833747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,1024,0.005695111221737332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,3072,0.013719999955760108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,7168,0.020622221959961783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,256,0.025238222546047632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,2048,0.01868799991077847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,65536,2.9063341352674694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,6144,0.22412888209025064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,16384,0.05578133132722643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,512,0.009551111194822524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,4096,0.020060444871584576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,1536,0.0123831108212471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,1024,0.0839395523071289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,16384,0.0351839992735121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,65536,0.13624266783396402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,7168,0.4249564541710748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,512,0.004882666799757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,1536,0.011149333582984077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,768,0.006568000134494569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,1536,0.0075777777367168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,65536,0.21171643998887804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,16384,9.650868733723959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,64,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,7168,2.9218701256646047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,16384,1.9649679395887587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,4096,0.017343110508388944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,65536,0.2638915644751655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,1536,0.013033777475357056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,64,0.006203555398517185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,3072,0.009313777916961247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,3072,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,128,0.11722400453355576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,1024,0.059691554970211454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,65536,0.43073778682284886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,16384,0.15972710980309382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,16384,0.31355733341640896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,2048,0.010870222416188983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,10240,0.0310533344745636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,4096,0.007154666715198093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,51200,1.1800951427883573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,4096,0.025617778301239014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,768,0.00598933340774642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,6144,0.875091552734375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,6144,0.3359031147427029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,32,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,16384,0.06208088662889269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,2048,0.5536070929633247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,64,0.0038568890757030914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,8192,0.421783102883233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,64,0.0196115556690428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,3584,0.01632000009218852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,5120,0.13594667116800943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,512,0.006189333481921091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,10240,0.07269599702623156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,3584,0.012357333468066322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,32,0.006526222245560751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,51200,2.0669146643744574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,512,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,32,0.008047111332416534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,7168,0.035066667530271746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,10240,0.00958933350112703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,256,0.004061333421203825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,10240,0.28805067804124623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,256,0.004044444610675176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,64,0.002793777734041214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,3584,0.23063556353251138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,64,0.014666666587193808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,5120,0.07292355431450738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,3072,0.008554666406578487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,3072,0.10130221976174249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,2048,0.015015999476114908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,512,0.006254222244024277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,3584,0.010984000232484607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,768,0.036933332681655884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,3072,0.02251822253068288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,7168,0.11115822527143691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,12288,0.16000000635782877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,32,0.17166577445136177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,8192,0.3400711218516032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,2560,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,5120,0.023599111371570166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,65536,0.07589244180255465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,512,0.007812444534566667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,3584,1.0324871275160048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,1536,0.03184266553984748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,1024,0.005033777819739448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,2560,0.008849778109126622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,1536,0.00776711106300354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,32,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,128,0.005688888745175467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,256,0.25125333997938365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,2048,0.04749066630999247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,2560,0.00602311102880372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,32,0.003198222153716617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,768,0.029112888707054987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,1024,0.01641155613793267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,32,0.005415110952324337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,3584,0.05725244681040446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,3072,0.011830222275522021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,3584,9.199132283528646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,256,0.031665778822369046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,256,0.012135999898115793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,7168,0.011844444606039258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,2048,0.010647999743620554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,1536,0.017475555340449016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,2560,0.01314933349688848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,3584,0.011538666983445486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,6144,0.12848888503180608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,768,0.009343110852771336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,7168,1.3264168633355033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,3584,0.010314666562610203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,6144,0.016020443704393175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,512,0.01647733317481147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,256,0.03715111149681939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,51200,0.19090489546457926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,7168,1.8142435285780165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,2560,0.0136808885468377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,5120,0.029398222764333088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,32,0.04357066750526428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,3072,0.015549333559142219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,32,0.005006222261322869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,512,0.014161777165200977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,3072,0.10346221923828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,12288,0.019387554791238572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,2560,0.0436462230152554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,10240,0.5701093143886989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,32,0.0036951113078329298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,2048,0.011457777685589261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,3584,0.06581066714392768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,51200,0.055679109361436635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,6144,0.03161066770553589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,64,0.04939377639028761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,65536,1.0133778254191081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,51200,0.18218844466739229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,7168,0.03405155407057868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,128,0.018191110756662156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,65536,6.154546525743272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,1536,0.013277333643701343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,8192,0.25312089920043945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,128,0.003145777723855443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,6144,0.026922666364245947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,768,0.013834666874673633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,3072,0.16981866624620226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,4096,0.03671377897262573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,3584,0.00983199973901113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,128,0.002806222273243798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,2048,0.0068764446510208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,2560,0.04467111163669162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,16384,0.008269333177142674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,2048,1.643140474955241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,64,0.00332266671790017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,12288,0.05376533336109585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,3584,0.017684444785118103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,512,0.018171555466122098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,2560,0.016972444123691983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,768,0.0035182221068276297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,12288,3.909776899549696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,65536,0.16460711426205105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,3584,0.069669332769182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,128,0.022600889205932617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,16384,0.17617155445946586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,5120,0.011222222612963783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,512,0.0177502218219969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,7168,0.025820444027582806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,6144,0.020397333635224234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,51200,0.7771920098198785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,10240,0.3329235447777642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,1024,0.34802132182651097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,128,0.008440000315507254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,12288,0.12159111764695908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,256,0.007860444486141205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,4096,0.1612266699473063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,3072,0.010499555203649731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,8192,0.028499556912316218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,3072,0.008030222521887885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,2048,0.8166960080464681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,3584,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,51200,0.2052639987733629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,6144,0.009964444571071202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,512,0.008768888811270395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,6144,0.2796764373779297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,2560,0.007706666986147563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,64,0.00816000004609426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,8192,0.06477866570154826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,51200,29.033050537109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,65536,3.5645217895507812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,128,0.016175111134847004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,1024,0.011571555501884885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,8192,0.013884444203641681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,1024,0.005700444595681296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,64,0.029919998513327703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,1024,0.014547554983033074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,4096,0.010496888723638324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,3584,0.11569066842397054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,16384,0.28883200221591526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,12288,0.03759200043148465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,128,0.008257778154479133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,8192,0.12793866793314615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,7168,0.02443644404411316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,3072,0.007700444095664554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,12288,0.01553066737122006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,10240,0.006872889068391588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,768,0.005872000008821487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,6144,0.4424826569027371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,8192,0.2536373403337267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,768,0.009815110928482479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,64,0.018717333674430847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,10240,0.03563288847605387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,6144,0.049845334556367665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,3584,0.008215111162927415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,51200,0.24920090039571127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,3072,0.015609777635998197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,12288,0.9777093463473849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,2048,0.09949155648549397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,3584,0.020608888732062448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,2560,0.08668888939751519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,3584,0.013199110825856527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,512,0.008028444316652086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,8192,0.018535999788178336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,1024,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,128,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,256,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,7168,0.011949333051840464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,65536,2.0685414208306208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,768,0.06614755259619819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,64,0.1722257799572415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,3072,0.010977778169843884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,128,0.01018133345577452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,5120,0.04170755545298258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,8192,0.052285333474477134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,10240,0.022687110635969374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,12288,0.19614577293395996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,1536,0.11898844771915013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,8192,0.13427200582292345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,8192,0.12377688619825576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,8192,0.015023110641373528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,2560,0.01633066601223416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,6144,0.015792888071801927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,7168,0.49976444244384766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,256,0.004080888711743885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,1024,0.013379555609491138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,7168,0.06307199928495619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,1024,0.011023110813564725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,7168,0.01441422270403968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,256,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,51200,0.04357422060436673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,6144,0.09675911400053237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,8192,0.06198933389451769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,768,0.1601075596279568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,51200,1.1589582231309679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,2048,0.012938666674825879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,10240,0.09078933133019342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,3584,1.5702515708075628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,1536,0.028024888700909082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,64,0.050683554675843984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,1536,0.009090666969617208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,10240,0.034881777233547635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,32,0.002846222163902389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,12288,1.6737795935736761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,16384,6.5021544562445746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,32,0.02896266513400608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,7168,0.023180445035298664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,32,0.021886222892337378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,65536,1.2378639645046658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,65536,0.39659643173217773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,12288,0.23154134220547148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,1024,0.007746666669845581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,6144,0.016033778587977093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,1536,0.0047973332305749255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,32,0.008610666626029545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,128,0.005106666849719154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,3584,0.03123377760251363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,2048,0.016333333320087857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,16384,0.39713690016004777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,64,0.0030097777230872046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,1536,0.009974222216341231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,65536,5.42940690782335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,2048,0.2338399887084961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,1024,0.01683911184469859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,3584,0.12398755550384521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,51200,0.9966373443603516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,5120,0.01603200038274129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,2048,0.020749333831999037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,51200,1.4717547098795574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,5120,0.02793688906563653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,10240,0.010288889209429422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,5120,0.01860000027550591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,8192,0.6320320235358344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,4096,0.06072711282306247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,512,0.0064328888224230866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,7168,0.009776888622177972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,256,0.006561777657932705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,51200,1.5751866234673395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,4096,2.5212995741102433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,10240,0.1617937750286526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,4096,0.009071110851234859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,1024,0.016252444850073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,10240,0.09965333011415269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,128,0.017550221747822232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,51200,1.5278239780002172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,256,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,7168,0.009860444400045607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,10240,0.012720000412729053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,32,0.008026666939258575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,512,0.012539555629094442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,51200,5.049595726860894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,16384,2.449728012084961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,256,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,4096,0.01088444474670622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,2560,0.014093332820468478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,1024,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,256,0.00645955569214291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,51200,0.09561866521835327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,768,0.008732444710201686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,3072,0.011267555256684622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,3072,0.016957332690556843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,4096,0.18732266955905488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,2560,0.017785777648289997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,128,0.020477333002620272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,8192,0.022108443909221228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,4096,0.009932444327407414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,512,0.027672888504134283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,6144,0.010344888601038191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,51200,0.5750160217285156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,4096,0.21761243873172334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,8192,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,16384,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,8192,0.012525333298577202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,4096,0.014969777729776172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,256,0.012329777909649743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,256,0.003395555540919304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,7168,0.23104000091552734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,2560,0.02269688910908169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,16384,0.29584977361891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,64,0.005150222116046482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,2048,0.011789333489206104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,64,0.003608889049953885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,16384,0.09928711255391438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,3584,0.016544888416926067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,16384,1.0886426501803927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,32,0.0041786668201287585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,51200,0.7153306537204318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,2560,0.01753866672515869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,2048,0.020443555381562974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,7168,0.09722666607962714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,1024,0.019108444452285767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,256,0.022802667485343084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,16384,0.11714577674865723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,65536,4.99491204155816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,512,0.0057297779454125305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,10240,0.02769511110252804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,4096,0.023601777023739282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,4096,0.10529777738783096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,51200,2.0999493069118924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,128,0.0064195555945237475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,8192,0.027612444427278306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,7168,1.4067618052164714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,2048,0.06480089161131117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,65536,0.1325955523384942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,256,0.0034506666577524612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,7168,0.07030666536755033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,5120,1.8842701382107203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,3072,0.011547555526097616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,3072,0.051742222574022084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,10240,0.03769777880774604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,32,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,2048,0.16869510544670951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,32,0.04040088918473985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,4096,0.12871289253234863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,1536,0.07082311312357585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,64,0.005386666705211003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,4096,0.011149333582984077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,2048,0.011588444312413534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,12288,0.02760444415940179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,10240,0.02437422176202138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,16384,0.021898667017618816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,32,0.02274844381544325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,768,0.007048000064161088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,51200,0.0698782205581665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,64,0.0053297777970631914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,12288,0.39243290159437394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,8192,0.04929600159327189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,512,0.003314666656984223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,65536,0.8428435325622559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,2048,0.011100444528791638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,10240,0.026877333720525105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,512,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,512,0.015992000699043274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,8192,0.058021333482530385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,128,0.005504888792832692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,2560,0.024825778272416856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,8192,0.008570666942331526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,3584,0.02181688944498698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,2560,0.00776533368561003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,2048,0.011354666617181567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,10240,0.034953776333067156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,65536,0.34612356291876895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,65536,0.5577662256028917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,8192,0.2638559871249729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,256,0.01421511173248291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,3072,0.4932870864868164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,2048,0.23321244451734754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,256,0.004496888981925117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,4096,0.009453333086437648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,3072,0.008460444708665213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,1536,0.004640000147951974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,1024,0.053611556688944496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,256,0.01731911136044396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,1024,0.006009777800904379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,128,0.00555288874440723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,768,0.038675556580225624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,3584,0.5294773313734267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,12288,0.06045155392752754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,256,0.003007111036115222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,7168,0.022871999277008906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,4096,0.010261333651012844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,1024,0.010778666370444827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,12288,0.5713484552171495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,1024,0.011047999891969891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,7168,0.0075288886825243635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,65536,0.11108977927101983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,3584,0.0625786648856269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,10240,0.01787911189926995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,12288,0.007853333320882585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,256,0.002918222298224767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,5120,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,16384,0.68777953253852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,2560,0.008222222328186035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,65536,0.3165208763546414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,2560,0.009376000199053023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,768,0.011170667078759937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,10240,0.006396444307433234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,3072,0.22254488203260633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,8192,0.10408977667490642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,10240,0.1233324474758572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,51200,0.10772533549202813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,256,0.00397155558069547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,3072,0.012619555824332766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,1536,0.0169831116994222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,512,0.00850311087237464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,1536,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,6144,0.031034668286641438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,4096,0.029812445243199665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,12288,0.14247377713521323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,4096,0.12559378147125244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,512,0.00812533332241906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,1536,0.05333422289954292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,5120,0.2897040049235026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,7168,0.018711111611790128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,256,0.004962666581074397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,5120,0.01443555619981554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,8192,0.21911466121673584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,2048,0.0074826669361856245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,3072,0.09389066696166992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,65536,0.07531822390026517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,12288,0.1097724437713623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,1024,0.3093991014692518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,8192,0.27323110898335773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,4096,0.07265066438251071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,128,0.003351111171974076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,256,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,6144,0.11533155706193711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,768,0.02836977773242527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,10240,0.2907608879937066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,128,0.004452444612979889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,12288,0.23232799106174043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,512,0.02353333267900679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,32,0.0029404444826973807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,128,0.0033493333806594214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,1024,0.003951999876234266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,51200,0.027106665902667578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,1536,0.016336000627941556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,51200,0.5146417617797852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,51200,0.41576443778143984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,10240,2.628397411770291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,768,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,5120,0.044963555203543774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,12288,0.04066666629579332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,2560,0.012832000023788877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,2048,0.04477689001295301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,2560,0.09202577670415242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,128,0.006703111032644908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,768,0.004004444513056013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,1536,0.00654666663871871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,3584,0.0075795559419526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,10240,7.1595408121744795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,8192,2.0501821306016708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,2560,0.8036657969156901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,1024,0.004848888764778773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,51200,0.3051431179046631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,128,0.0071937781241205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,12288,0.25787199868096244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,3072,0.01644355555375417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,3584,0.07687911060121325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,128,0.2850124571058485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,5120,0.09373955594168769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,10240,0.41635645760430234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,2048,0.012534221841229333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,256,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,3072,0.4404462178548177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,8192,0.059260441197289355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,512,0.006066666709052191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,65536,0.013422222601042854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,10240,0.10317244132359822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,51200,0.009268444445398118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,65536,0.5246195793151855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,5120,0.01124533348613315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,512,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,3072,0.010975999964608086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,256,0.003084444544381566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,32,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,1536,0.02225688927703434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,1536,0.12688355975680882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,128,0.011662222445011139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,768,0.031353778309292264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,2048,0.013204444613721637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,32,0.005023111071851519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,64,0.004969777746333016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,51200,0.45176267623901367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,3584,0.1813217798868815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,4096,0.008223999705579545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,8192,0.6082737710740831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,128,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,10240,0.07192089160283406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,51200,0.024645333488782246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,65536,0.21089333958095977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,3584,0.4494622018602159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,12288,0.03681333197487725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,768,0.00398577791121271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,1024,0.004570666700601578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,128,0.09902933571073745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,16384,0.34200178252326113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,3584,0.025832888152864244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,10240,0.018339556124475267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,12288,0.04782933327886793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,3584,0.01828177769978841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,12288,0.024088889360427856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,1536,0.010888888604111142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,32,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,512,0.008074666890833113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,256,0.0155102229780621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,1536,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,256,0.01735733283890618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,6144,0.028691556718614366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,128,0.01368711143732071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,128,0.008157333566082848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,6144,0.06172000037299263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,1024,0.0057955553962124725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,65536,0.08819466829299927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,128,0.0028373334142896864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,1024,0.011110222174061669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,32,0.016608888904253643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,16384,0.06179110871420967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,64,0.005103999955786599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,256,0.11263644695281982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,64,0.06826044453514947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,5120,0.0070426662762959795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,65536,0.22362221611870658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,1536,0.018761777215533786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,64,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,128,0.003125333330697484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,2560,0.11967466937171088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,10240,0.0669635534286499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,3072,0.2981315453847249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,16384,0.2153528928756714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,4096,0.014001778430408902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,2560,0.007939555578761632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,6144,0.015882665912310284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,256,0.008201777935028076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,16384,0.034499555826187134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,2048,0.05397511190838284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,8192,0.06241333484649658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,3584,0.19988089137607148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,768,0.23481776979234484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,32,0.0034453332838084963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,2560,0.03354933195643955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,7168,0.39408267868889707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,12288,0.00664533343580034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,5120,0.019115555617544387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,65536,0.24926132626003691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,65536,0.2897084554036458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,7168,0.06264711088604398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,8192,0.06964977582295735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,8192,0.01884355478816562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,7168,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,4096,0.008378666308191087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,7168,0.019869332512219746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,65536,0.33249423238966197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,7168,0.9106391270955404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,512,0.007074666519959767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,768,0.004337777694066365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,51200,0.06431733237372504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,128,0.004050666673315896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,1536,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,4096,0.012660444610648684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,3072,0.008240000241332585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,512,0.00598933340774642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,2048,0.014364444547229342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,4096,0.09136444330215454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,12288,0.04749333196216159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,4096,0.009425777528021071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,65536,0.021469333105617102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,10240,0.32443377706739635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,1024,0.004482666651407878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,7168,0.03137777911292182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,16384,0.04452444447411431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,2048,0.015414221419228448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,7168,0.26803021960788304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,512,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,128,0.0069653333889113525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,51200,0.013861333330472311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,8192,0.014929778046078153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,3584,0.30453599823845756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,128,0.008151111503442129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,1024,0.004319111092223061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,768,0.018182223041852314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,64,0.04193244377772013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,768,0.006467555546098285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,768,0.004045333299371931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,7168,1.7278951009114583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,128,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,51200,0.10451555252075195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,51200,0.5623102188110352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,256,0.0032924444725116095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,2560,0.007327111230956183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,64,0.04158933295143975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,256,0.0034835553831524322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,65536,0.11368444230821396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,256,0.004477333277463913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,51200,8.779588487413195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,1536,0.02676800058947669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,16384,0.19559021790822348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,128,0.003933333274390963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,6144,0.015464888678656684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,1024,0.005420444326268301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,256,0.31967467731899685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,51200,0.06964800092909071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,2048,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,51200,0.3159484333462185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,12288,0.216139554977417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,256,0.003306666596068276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,1536,0.014769777655601501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,2048,0.006907555378145642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,5120,0.11017777522404988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,8192,0.02362577782736884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,64,0.00434755575325754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,10240,0.018005333013004728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,768,0.003967111309369405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,3072,0.1709208885828654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,16384,0.022050667140218947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,5120,0.11224888430701362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,12288,0.154950221379598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,256,0.005432888865470886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,65536,0.10616088575787014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,3072,0.010488889283604093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,51200,0.14981955952114528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,512,0.004986666556861666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,8192,0.022030222747060988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,5120,0.010073777702119615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,10240,0.04876533481809828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,3584,0.011708444191349877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,2560,0.007072889142566257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,4096,0.011796444654464722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,2048,0.08148266871770223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,64,0.019878221882714164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,51200,7.579728020562066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,10240,0.014974221587181091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,512,0.16319110658433703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,6144,0.02619733413060506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,768,0.010690666735172272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,2560,0.006762666834725275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,3584,0.019323555959595572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,12288,0.1311404440138075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,6144,2.3939200507269964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,3072,1.6824817657470703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,256,0.0042239998777707415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,7168,0.051503111918767296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,512,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,256,0.04587466518084208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,10240,0.06841066810819837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,10240,0.48936976326836484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,51200,0.8286462359958224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,16384,0.3977324432796902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,8192,1.4146657519870336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,1536,0.008309333688682979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,16384,0.03646577729119195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,5120,0.16918578412797716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,3584,0.05994844436645508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,512,0.012191111014948951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,16384,0.019069333871205647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,32,0.007258666886223688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,51200,4.418454064263238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,16384,0.0387386679649353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,768,0.00602311102880372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,16384,0.04760800136460198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,2560,0.01848444508181678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,512,0.031776888502968684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,16384,0.1559928920533922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,12288,0.08214044570922852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,51200,0.11720177862379287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,12288,0.03810844487614102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,10240,0.42441511154174805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,2560,0.014279999666743808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,7168,0.01926577753490872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,16384,0.07726222276687622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,1536,0.08829689025878906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,128,0.007814221911960179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,12288,0.017535999417304993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,2560,0.19433066580030653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,3584,0.16841777165730795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,2048,0.008027555214034187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,10240,0.5090728865729438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,1536,0.07509155405892266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,4096,0.02384800049993727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,6144,0.10882311397128636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,256,0.003097777772280905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,65536,0.1315351062350803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,3072,0.016891555653678048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,5120,0.01761511133776771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,768,0.018743111027611625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,64,0.005249777601824866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,65536,3.2340977986653647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,65536,1.1223831176757812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,12288,0.23795554372999403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,7168,0.016944888565275405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,6144,0.015232889188660515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,10240,0.046803557210498385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,2048,0.008747555315494537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,256,0.016034666034910414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,2560,0.31175110075208873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,2560,0.05079377690951029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,128,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,5120,0.04955822229385376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,6144,0.019268444842762418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,256,0.012958221965365939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,12288,0.03132799930042691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,16384,0.6193635728624132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,3584,0.046853333711624146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,2048,0.030910223722457886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,1536,0.010089777410030365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,768,0.013255111045307584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,128,0.014197332991494073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,3072,0.01518311103185018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,128,0.007843555675612556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,3072,0.0539973345067766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,64,0.0075599998235702515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,3072,0.02637422250376807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,6144,0.5258577664693197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,12288,43.930945502387154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,3584,0.03369244601991441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,64,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,12288,0.15887999534606934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,4096,0.0655662218729655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,2048,0.011145778000354767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,1024,0.00608800020482805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,256,0.004119111018048393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,5120,0.02488800055450863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,2048,0.04118400149875217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,128,0.17413777775234648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,16384,0.0799822211265564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,3072,0.28916088740030926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,12288,0.023411555422676936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,8192,1.3140577740139432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,128,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,768,0.011607111328177981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,2560,0.015075556106037565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,128,0.106532441245185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,8192,0.02426311042573717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,1024,0.7038906415303549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,256,0.003605333467324575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,4096,0.1066017746925354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,2048,0.029318223396937054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,1536,0.019301333361201815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,4096,0.07109066512849596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,2048,0.10346311330795288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,256,0.003105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,10240,0.04145155681504144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,16384,0.2867457866668701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,6144,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,7168,0.14413955476548937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,256,0.002862222285734283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,1536,0.004707555390066571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,8192,0.013632000320487551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,7168,0.028941333293914795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,2560,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,2560,0.0741786691877577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,512,0.008447999755541483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,12288,0.04362666606903076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,65536,0.8037661976284451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,6144,0.33717065387301975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,51200,0.008059555457697975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,64,0.002620444529586368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,768,0.0063031112982167145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,51200,0.3152711126539442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,3584,0.007096000015735626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,1024,0.005445333404673471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,3072,1.2287875281439888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,51200,0.11677777767181396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,3072,0.018358222312397428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,12288,0.5663475460476345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,128,0.003731555408901639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,512,0.36306044790479874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,32,0.004779555731349521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,3584,1.1528844833374023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,3072,0.01128088848458396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,6144,0.15372888247172037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,12288,0.028239111105600994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,8192,0.04454044500986735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,4096,0.0388408899307251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,256,0.005487111293607288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,51200,0.09187377823723687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,16384,0.1040453314781189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,4096,0.030384888251622517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,512,0.03233688738611009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,256,0.01793599956565433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,4096,0.11164977815416123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,12288,0.07013689147101508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,1024,0.009214222431182861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,7168,0.00740533322095871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,10240,0.05563733312818739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,6144,0.013961777091026306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,7168,0.010944888823562197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,2048,0.1260719961590237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,16384,0.05545511179500156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,16384,0.7749590873718262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,32,0.0038213332494099936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,8192,0.017196445001496207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,5120,1.7352897855970595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,2048,0.2853848934173584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,7168,0.012995555169052549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,7168,0.0920097761683994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,3584,0.1096915536456638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,6144,0.02558311157756382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,128,0.12401155630747478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,2560,0.055454221036699086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,4096,0.21235022279951307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,4096,0.08394755257500543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,51200,0.1368524498409695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,10240,4.231718275282118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,2560,0.04306133257018196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,16384,0.02089599933889177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,3072,0.007107555866241455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,51200,0.13540355364481607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,3584,0.1206479999754164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,51200,2.0891138712565103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,8192,0.2733164363437229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,3072,0.2239289018842909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,51200,0.1314106649822659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,12288,0.42261777983771426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,12288,0.21834222475687662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,16384,0.035804443889194064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,3584,0.05548000004556444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,1024,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,1024,0.009293333523803288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,16384,0.20607288678487143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,1536,0.015844444433848064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,768,0.0036391110883818734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,5120,0.012138666378127204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,4096,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,16384,0.022258667482270136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,3584,0.016806221670574613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,3584,0.007809778054555257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,6144,0.04933155576388041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,256,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,768,0.018754666050275166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,3072,0.023639111055268183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,2048,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,16384,0.051623112625545926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,3584,0.12730844815572104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,1536,0.09099644422531128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,768,0.005404444618357553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,65536,0.160316440794203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,64,0.0031911111954185697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,10240,0.0471111105548011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,1024,0.007119999991522894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,1536,0.007206222249401941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,8192,0.05557777484258016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,5120,0.01703555550840166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,8192,0.2390373282962375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,7168,0.15490755769941542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,6144,0.020928889513015747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,5120,0.04698044392797682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,5120,0.1512426667743259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,2048,0.013376889129479727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,51200,1.345615069071452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,16384,0.14240622520446777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,3072,0.09090844127866958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,768,0.31612179014417857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,512,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,51200,0.10032888915803696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,8192,0.5232942369249132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,8192,0.03428355521625943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,128,0.005314666777849197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,8192,0.04100799891683791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,65536,0.5129449102613661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,65536,0.0605857769648234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,1024,0.04478755593299866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,2048,0.00815644446346495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,64,0.0026586666289303037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,32,0.00315022220214208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,768,0.004960888789759742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,3072,0.02732533382044898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,512,0.00618488879667388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,5120,0.010175111393133799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,768,0.010079111489984725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,32,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,32,0.00519466648499171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,16384,0.11005955272250706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,768,0.004661333229806689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,32,0.0027822222974565295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,3584,0.012109333442317115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,1536,0.032997333341174655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,128,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,256,0.007627555893527136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,1024,0.008652444514963362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,10240,0.5450124210781521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,256,0.18644444147745767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,2048,0.10349244541592068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,1536,0.02402488887310028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,32,0.0058044443527857465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,1536,0.056919998592800565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,3072,0.014388443695174323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,12288,3.5264320373535156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,12288,0.3684542179107666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,7168,0.029116445117526587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,5120,0.2055706712934706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,1536,0.009544889132181803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,65536,1.7644062042236328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,7168,0.3423004415300157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,3584,0.5716213120354546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,6144,0.01613066593805949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,256,0.011123555401961008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,51200,0.17951377232869467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,128,0.0042071110672420925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,6144,0.02737866673204634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,10240,0.0454364452097151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,512,0.019510222805870902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,2048,0.01589866644806332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,8192,0.04911288950178358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,768,0.05787822273042467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,6144,0.016893333858913846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,7168,0.16532088650597468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,512,0.01716888944307963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,32,0.00535111129283905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,51200,0.445093313852946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,128,0.0047733332547876565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,7168,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,128,0.010984000232484607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,6144,0.31848178969489205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,6144,0.01793688866827223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,768,0.004379555583000183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,256,0.020443555381562974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,64,0.012034667034943899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,128,0.008700444466537898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,6144,0.019703111714786954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,5120,0.011403555671374003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,768,0.006684444430801604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,32,0.00742222203148736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,512,0.007376888559924231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,2048,0.11598755253685845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,3072,0.08225599924723308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,64,0.0027217778066794076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,3072,0.025032889511850145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,768,0.00961599995692571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,128,0.005383111122581694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,1024,0.00979466653532452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,1024,0.008926221893893348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,10240,0.013815110756291283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,64,0.0032275555034478507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,2560,0.022592888938056097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,32,0.21255644162495932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,1536,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,2560,0.018127111925019156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,256,0.009661333428488838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,32,0.01054044481780794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,2048,0.037040889263153076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,1024,0.009482666850090027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,32,0.005096000101831225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,32,0.008735111190213097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,12288,0.01664177742269304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,2560,0.06674577792485555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,8192,0.018059555027219985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,1536,0.0063368889192740125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,16384,0.18400711483425566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,4096,0.011666666302416058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,1536,0.0064791109826829694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,12288,0.8252293268839518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,5120,0.14565511544545492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,1536,0.0063786668082078295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,10240,0.7668844328986274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,32,0.0032586666444937387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,2560,0.01442311041884952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,51200,0.2444737752278646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,12288,0.02008888953261905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,3072,0.0063324446479479475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,256,0.004013333469629288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,65536,0.183789332707723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,6144,0.1331662204530504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,3584,0.012651555240154266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,768,0.007124444676770105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,12288,0.07718666394551595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,10240,0.037089778317345515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,3584,0.13677688439687094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,3584,0.0336773329310947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,1024,0.0052151112920708126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,12288,0.0492524438434177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,1536,0.2622613377041287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,5120,0.028976887464523315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,5120,0.00885511106914944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,8192,0.3100950982835558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,16384,0.6948417557610406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,3072,0.008533333738644918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,10240,0.01835644410716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,3072,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,12288,0.036222222778532244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,16384,3.854651557074653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,1536,0.03423555692036947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,1024,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,3584,0.02924088968171014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,51200,0.01794399983353085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,3072,0.022112000319692824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,3584,0.1299955579969618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,4096,0.011164444188276926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,8192,0.5552106433444554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,2048,0.006052444378534953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,12288,0.01991911066903008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,4096,0.044221334987216525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,64,0.013745778136783175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,512,0.009691555466916826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,12288,8.76969485812717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,1536,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,768,2.833200878567166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,10240,0.10173955228593613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,512,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,512,0.004304000073009067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,2048,0.009402666654851701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,4096,0.014225777652528552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,2048,0.01000711073478063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,64,0.00637155564294921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,12288,0.021930666433440313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,51200,0.11061777671178182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,128,0.0028880000528362063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,768,0.010375111467308467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,6144,0.010092444717884064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,512,0.00798044436507755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,10240,0.1655422184202406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,65536,2.011470158894857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,128,0.013778666655222574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,12288,0.008424888882372115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,6144,0.01871288816134135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,1024,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,768,0.03486311104562547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,65536,0.8253288798862033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,12288,0.315465768178304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,10240,0.024939555260870192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,5120,0.021363554729355708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,64,0.003093333293994268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,51200,0.05335022343529595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,2048,0.015470221638679504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,128,0.003119999956753519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,7168,0.013839110732078552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,2560,0.016401777664820354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,10240,0.0211173329088423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,5120,0.01770400007565816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,2048,0.10240088568793403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,128,0.002752888947725296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,1536,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,7168,0.14602222707536486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,1024,0.0040755557517210645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,1024,0.006769777586062749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,5120,0.04455644554562039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,16384,1.6327368418375652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,32,0.0034044444974925783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,128,0.006299555715587404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,512,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,51200,0.14937155776553684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,2560,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,1536,0.053048001395331494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,1024,0.00812888890504837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,12288,0.015519110692871941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,65536,0.08073688877953424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,10240,0.035628444618648954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,2048,0.018768888380792405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,6144,0.006110222389300664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,8192,0.01366133325629764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,2560,0.007445333732499017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,1024,0.010652444428867765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,1024,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,51200,0.027986665566762287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,768,0.009328000247478485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,10240,0.013254221942689685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,32,0.018813333577579923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,5120,0.008984000318580205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,2048,0.038880000511805214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,7168,0.18082133928934732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,65536,1.7840603722466364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,3584,0.02994133366478814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,7168,0.047839999198913574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,2560,0.013736888766288757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,2048,0.0068764446510208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,51200,0.055600888199276395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,1024,0.0581173300743103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,2560,0.03739555676778158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,768,0.0076044441925154785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,256,0.022987556126382615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,4096,0.199562668800354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,1024,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,51200,0.2528791162702772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,512,0.0037413334680928122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,768,0.012824888858530255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,1024,0.009840889109505547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,12288,8.709522671169704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,512,0.006308444258239534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,256,0.00675644435816341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,768,0.007349333001507654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,4096,0.03678933448261685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,5120,0.0074888889988263445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,768,0.015655999382336933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,3072,0.013693333499961428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,256,0.005160000175237656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,10240,0.023919110496838886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,64,0.006735111276308696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,16384,0.6279884444342719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,3072,0.010658666491508484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,7168,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,65536,0.12715289327833387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,3584,0.01146399974822998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,32,0.0032008888406885993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,1024,0.021212443709373474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,1536,0.005411555369695027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,2560,0.009290667043791877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,6144,0.07796266343858507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,3072,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,4096,0.19575822353363037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,6144,0.020936000678274367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,3072,0.1538053353627523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,256,0.004530666602982415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,128,0.0029893333299292457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,16384,0.06981244352128771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,2560,1.2293031480577257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,1536,0.012794666820102267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,64,0.005961777849329843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,65536,3.792466481526693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,1024,0.01311288856797748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,2048,1.761799070570204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,768,0.020592000749376085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,768,0.011152000062995486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,8192,0.06953599717881945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,65536,0.16253244876861572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,3584,0.10441155566109551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,128,0.016183111402723525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,65536,4.566034528944227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,256,0.003685333248641756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,2560,0.02780533168050978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,4096,0.011071110765139261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,65536,1.0613501866658528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,4096,0.00998933364947637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,512,0.048715555005603366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,64,0.003120888852410846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,6144,0.09694221946928237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,2560,0.07018311156166925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,256,0.004510222209824456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,5120,0.23286843299865723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,32,0.015343111422326831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,256,0.004226666771703296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,16384,0.3154711193508572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,1024,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,4096,0.007083555890454187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,65536,0.08003199762768216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,8192,0.27180443869696724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,12288,0.6483670870463053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,3584,0.0207324441936281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,3584,0.013705777625242868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,10240,0.03418488966094123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,10240,2.827151192559136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,8192,0.017493334081437852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,32,0.018193778064515855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,6144,0.06614933411280315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,128,0.008375999828179678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,16384,0.09370311101277669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,2560,0.026584888497988384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,128,0.01648622254530589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,16384,0.08255910873413086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,65536,0.3076106707255046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,3584,0.38910844590928817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,32,0.02312444481584761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,1024,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,4096,0.10702755716111924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,512,0.028892444239722356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,512,0.0075457774930530125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,7168,0.03154222170511881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,4096,0.01889244384235806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,2560,0.016747555798954435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,12288,0.5641706784566244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,2048,0.1988755597008599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,51200,1.7279404534233942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,8192,0.09274755583869086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,512,0.02271822260485755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,16384,0.9934470918443469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,512,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,7168,0.1441297796037462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,2560,0.011938666303952536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,128,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,16384,0.5696124500698513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,8192,0.17804266346825492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,12288,1.1737724939982097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,512,0.025246222813924152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,12288,2.813803566826714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,128,0.004214222232500712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,65536,1.098101298014323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,32,0.009588444398509132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,51200,2.5762479570176864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,51200,8.797900729709202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,768,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,2560,0.02404266595840454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,6144,0.0750817788971795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,128,0.07345955239401923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,65536,0.4078542126549615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,256,0.00309955556359556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,32,0.005464000006516774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,65536,0.36701512336730957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,2048,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,12288,1.3543787002563477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,256,0.005904888941182031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,65536,0.09040266937679714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,51200,0.36188356081644696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,8192,0.017303110824690927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,768,0.003932444585694207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,2048,0.06850044594870673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,64,0.006074666563007567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,512,0.003957333250178231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,7168,0.00996888925631841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,6144,0.025120000044504803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,1536,0.01886311173439026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,3072,0.04721777637799581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,12288,0.451307561662462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,8192,0.019346666004922655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,4096,0.023702222439977858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,16384,0.021012443635198805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,4096,0.01612622208065457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,16384,1.4938827090793187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,16384,7.1560863918728295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,4096,0.058697779973347984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,16384,0.11923110485076904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,512,0.03358666764365302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,10240,0.1293813387552897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,10240,0.007720888488822513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,6144,0.14049600230322942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,32,0.013385777672131857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,65536,0.11155377493964301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,16384,0.038619554705089994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,7168,0.029309332370758057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,8192,0.03064089020093282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,10240,0.0158897770775689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,4096,0.00867555538813273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,12288,4.9447102016872835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,3072,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,1536,0.12885155942704943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,32,0.003208000005947219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,7168,0.27536532613966197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,7168,2.7166400485568576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,4096,0.014891554911931356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,10240,0.029603554142846003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,3072,0.14481955104404026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,12288,0.2097173399395413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,2560,0.43281955189175075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,2048,0.008216888540320927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,65536,0.11475377612643772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,1024,0.008991111483838823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,768,0.003806222230195999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,256,0.005537777725193236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,8192,0.012254222399658628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,3584,0.011767999993430244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,256,0.004291555533806483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,2048,0.036158220635520086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,16384,0.030601776308483545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,6144,0.10944533348083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,65536,49.22637939453125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,3072,0.1281715499030219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,10240,4.4581014845106335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,2560,0.009220444493823582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,32,0.08326399988598294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,64,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,768,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,2560,0.019525332583321463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,2560,0.015991111596425373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,3072,0.012186666329701742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,1536,0.008060444560315874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,1536,0.010400888820489248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,51200,7.682037353515625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,16384,2.5101502736409507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,3584,0.014984000060293408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,51200,0.17084977361891004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,1024,0.006541333264774746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,51200,0.1933671103583442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,10240,0.021506667137145996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,6144,0.3048231071896023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,1536,0.018227555685573153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,1024,0.018578666779730055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,7168,0.023082666926913794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,16384,0.014841778410805596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,256,0.0032106666929192017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,5120,0.012320888539155325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,3072,0.11054666837056477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,10240,0.023670223024156358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,5120,0.03490666548411051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,4096,0.010087110930018956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,65536,0.10670133431752522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,3072,0.020599111914634705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,3584,0.024784000383483037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,256,0.009434666898515489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,128,0.014018666413095264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,1024,0.003956444561481476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,65536,2.865410698784722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,768,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,65536,0.010392000277837118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,2560,0.04088888896836175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,512,0.029779556724760268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,10240,0.02696711156103346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,256,0.007322666545708974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,5120,0.09345155292087132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,2560,0.020614221692085266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,2048,0.015691555208630033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,7168,0.0795324444770813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,2560,0.31338132752312553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,512,0.01720355616675483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,32,0.003318222239613533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,8192,0.015824889143308003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,3584,0.012328888807031842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,12288,0.00889066689544254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,512,0.20313510629865858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,1024,0.02163822286658817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,512,0.005689777847793367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,12288,0.1542337735493978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,51200,0.4614577823215061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,51200,0.029884444342719182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,16384,3.3996995290120444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,12288,0.25518398814731175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,51200,0.13801777362823486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,2560,0.027705777022573683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,3072,0.00741155528359943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,6144,0.2070711056391398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,2560,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,2048,0.0086666668454806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,512,0.021390222840838965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,768,0.009482666850090027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,12288,0.39424888292948407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,12288,0.31805067592196995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,10240,0.14762489000956217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,65536,4.417355431450738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,256,0.00518844442235099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,8192,0.1401404407289293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,2560,0.006270222365856171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,10240,0.013152889079517789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,12288,0.009995555712117089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,16384,0.08976355526182388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,5120,0.019161777363883126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,64,0.006248888870080312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,3072,0.026563555002212524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,6144,0.04223733478122287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,6144,0.2677991125318739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,7168,0.16455111238691542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,1024,0.015631111131774057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,256,0.005390222287840313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,2048,0.1616106695599026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,65536,0.14739289548661974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,8192,0.042562666866514415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,32,0.03488711184925503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,64,0.0026364444444576898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,1536,0.0064328888224230866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,3072,1.0727679994371202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,8192,0.3683368894788954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,12288,0.03243911266326904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,32,0.002777777819169892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,1024,0.015951999359660678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,16384,9.507528516981337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,10240,0.03170311119821336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,1024,0.684474680158827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,3072,0.20588800642225477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,6144,0.012303111453851065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,768,0.09062577618492974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,3072,0.630317317114936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,4096,0.011491555306646558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,7168,0.23086934619479707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,65536,1.8649395836724176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,768,0.03495466709136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,8192,0.2066213289896647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,8192,0.0912826657295227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,32,0.007975111405054728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,51200,0.17255732748243544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,8192,0.02299022177855174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,3584,0.007620444728268518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,7168,0.01531377765867445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,8192,0.05422844489415487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,51200,0.2783244450887044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,10240,0.04077422287729051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,65536,0.0329786671532525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,4096,0.011959999799728394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,8192,0.09093244208229913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,8192,0.017487110363112558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,1536,0.007262222468852997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,128,0.0033884443756606844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,5120,0.04581244455443489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,51200,0.15426133738623724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,16384,0.23723112212287056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,768,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,2048,0.16391289234161377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,1536,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,16384,0.05166133244832357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,5120,0.022135110365019903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,2560,0.007401777638329401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,2048,0.017408889200952318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,12288,0.07408977879418267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,65536,0.21929689248402914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,8192,0.022914666268560622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,1024,0.007000889215204451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,4096,0.014941333068741692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,128,0.009988444546858469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,4096,0.023005333211686876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,64,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,2560,0.009492444495360056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,65536,2.3082489437527127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,256,0.018328888548745047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,16384,6.911641862657334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,64,0.009370666411187913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,2048,0.012210666305489011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,4096,0.015575110912322998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,6144,2.45053948296441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,10240,0.1881786717308892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,51200,2.148743099636502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,256,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,2048,0.13086488511827257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,65536,2.278270297580295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,8192,0.041856888267729014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,32,0.0542719993326399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,768,0.011617778076065911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,2048,0.021869333253966436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,6144,0.02132533325089349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,512,0.14076799816555446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,3584,0.017270222306251526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,2560,0.007159999675220913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,65536,0.4910737673441569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,768,0.015268445014953613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,1536,0.036891556448406644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,2560,0.12797599368625218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,3072,0.007650666766696506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,128,0.013243555194801755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,5120,0.012988444831636218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,768,0.02070399953259362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,512,0.040842665566338435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,2560,0.009320889082219865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,10240,0.01696266730626424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,128,0.008303110798199972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,3584,0.12287110752529568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,256,0.11769778198666042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,3584,0.05969688627454969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,65536,0.2348844475216336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,12288,0.0449848903550042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,3072,0.08282666736178927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,10240,0.15373333295186362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,5120,0.010951110886202918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,1536,0.010657777388890585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,128,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,6144,0.06881333059734769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,3584,0.08055733309851752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,2560,0.01856355534659492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,16384,0.049878223074807056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,5120,0.01219555570019616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,51200,43.895185682508675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,4096,0.03525866733656989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,512,0.06065066655476888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,3072,0.011404444773991903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,2048,0.009726222190592024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,2048,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,65536,0.2988177670372857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,256,0.011933333343929715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,128,0.007119110888904995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,2560,0.009478222164842818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,65536,17.681705050998264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,10240,0.01702577703528934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,768,0.00942666663063897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,5120,0.0939377744992574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,8192,0.025586666332350835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,128,0.24421956804063585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,65536,25.000995212131077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,7168,0.12188177638583714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,768,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,128,0.010867555936177572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,512,0.008376888930797577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,2560,0.013277333643701343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,256,0.009297777381208207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,51200,0.15714755323198107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,1536,0.011887999872366587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,2560,0.016207110550668504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,4096,0.008995555341243744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,5120,0.009275555610656738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,256,0.007382222347789341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,1536,0.007162666983074612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,128,0.01905955539809333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,3584,0.013187555803192986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,2048,0.010972444381978778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,32,0.005028444445795483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,256,0.004580444345871608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,7168,0.018018666240904067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,7168,0.02702577743265364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,2048,0.006686222222116258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,1536,0.007716444631417592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,32,0.0031635556370019913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,768,0.0063697778516345555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,3584,0.0333262218369378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,6144,4.050969865587023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,5120,0.015156444576051501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,2048,0.010064000056849586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,2560,0.009447111023796929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,4096,0.01350400017367469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,65536,0.0820888876914978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,4096,0.05317599905861748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,128,0.14584978421529135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,32,0.0029173334025674393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,51200,0.7117724418640137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,5120,0.01479822231663598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,256,0.0037671112351947357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,64,0.0025742221623659134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,1536,0.014365333649847241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,4096,0.032666666640175715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,3584,0.008107555409272512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,3072,0.018558222386572096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,2560,0.514335102505154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,5120,0.3909946547614203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,12288,0.2162711090511746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,768,0.009138666921191746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,51200,5.522991180419922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,2048,0.11353777514563666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,32,0.017456000049908955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,65536,0.2213182184431288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,32,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,12288,0.04526755544874403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,6144,0.0192239996459749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,2048,0.021792888641357422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,256,0.00852177788813909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,3584,0.010194666683673859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,256,0.008630221916569604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,12288,0.07450933588875665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,512,0.00545155546731419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,8192,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,4096,0.045421332120895386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,256,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,2560,0.0162764439980189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,256,0.007800888684060838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,512,0.011271110839313932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,65536,0.12762399514516196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,2560,0.016388444436921012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,6144,0.7166168954637316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,7168,0.009224889179070791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,51200,0.2890373335944282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,51200,0.08330221970876057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,51200,0.17358044783274332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,12288,0.016181333197487723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,10240,0.3607564502292209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,1024,0.006154666758245892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,768,0.04367555512322319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,512,0.005703111075692707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,6144,0.014926221635606555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,3072,0.007240888973077138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,3584,0.008522666990756989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,3072,0.007452444069915348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,10240,0.019155555301242404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,8192,0.1633217732111613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,2560,0.008062221937709385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,512,0.09656000137329102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,5120,0.021371554997232225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,8192,0.27301155196295845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,32,0.011008000208271874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,12288,1.127097765604655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,12288,0.07407377825842963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,8192,0.38559378517998594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,3584,0.11356355084313287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,256,0.023261333505312603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,65536,1.193991978963216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,256,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,12288,0.009343110852771336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,65536,0.13992710908253989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,51200,1.6405760447184246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,10240,0.1309600008858575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,16384,0.06216444571812948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,3072,0.012576888832781049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,64,0.0037768888804647657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,768,0.12341777483622234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,1024,0.28155822224087185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,2048,0.018371555540296767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,16384,65536,1.6447306738959417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,7168,0.09406399726867676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,512,0.0051857779423395796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,3584,0.015471110741297403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,1024,0.006266666783226862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,128,0.004973333328962326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,2048,0.140301333533393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,512,0.021993777818149988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,51200,1.7675529056125219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,32,0.005712000032265981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,128,0.004849777867396672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,256,0.007149332927332983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,1536,0.00546488869521353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,10240,0.01518311103185018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,4096,0.02516799999607934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,512,0.06871733400556776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,6144,0.0331928895579444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,3072,0.03882222374280294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,32,0.10860266950395371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,51200,1.567184872097439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,1536,0.029932445949978296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,8192,0.43943378660413956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,7168,0.0313351121213701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,51200,0.16868089305029976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,51200,0.9083644019232856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,65536,0.14035911030239528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,512,0.0037155557009908888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,1536,0.0047457776963710785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,51200,0.30777422587076825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,5120,0.015235554840829639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,1024,0.007112888826264276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,2048,0.020619556307792664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,3584,0.02832088867823283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,2560,0.07052355342441134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,7168,0.2300915453169081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,12288,0.040008889304267034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,5120,0.012492444780137805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,16384,1.0841218100653756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,512,0.005864000154866113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,65536,0.25760798984103733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,2560,0.020076443751653034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,128,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,2560,0.009728000395827824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,65536,5.2054799397786455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,3584,0.029025776518715754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,2048,0.02070399953259362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,64,0.009151111046473185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,4096,0.04866133464707268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,2560,0.024519110719362896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,256,0.044714666075176664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,5120,0.012870222330093384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,3072,0.05870755513509115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,12288,0.06173955731921726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,256,0.010875555376211802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,256,0.005443555613358815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,64,0.0026853332916895547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,1536,0.0047253333032131195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,51200,0.06672444608476427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,6144,0.10715199841393365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,65536,0.1866977744632297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,3072,0.02237866653336419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,5120,0.009567110902733272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,256,0.005045333256324132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,1024,0.005856888989607493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,51200,0.01144266708029641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,10240,1.0325306786431205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,64,0.00498488876554701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,768,0.04621155394448174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,1024,0.009148444566461775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,3584,0.01276088919904497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,2560,0.01903733279969957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,4096,0.011152889165613385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,128,0.03554933269818624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,6144,0.03779110974735684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,6144,0.00720355576939053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,768,0.0627688897980584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,512,0.011557333171367645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,10240,0.01499288943078783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,128,0.005753777921199799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,5120,0.01793333391348521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,4096,0.012873777912722694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,16384,0.04126133190260993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,128,0.004491555607981152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,51200,11.521310594346788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,256,0.0064186664919058485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,3072,0.011319110790888468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,65536,0.19615999857584634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,16384,0.061758220195770264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,512,0.021158221695158217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,2560,0.165065778626336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,256,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,4096,0.021182222498787776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,512,0.024314666787783306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,2048,0.08976711167229547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,1024,0.1488986677593655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,10240,0.33297599686516655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,16384,0.14384710788726807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,8192,0.03422133458985223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,768,0.003624888757864634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,10240,0.08201422293980916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,128,0.00544177782204416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,32,0.002577777744995223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,512,0.00387999994887246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,8192,0.014067555467287699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,1536,0.011704000333944956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,256,0.00572533326016532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,128,0.0027635554886526535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,10240,0.007311111523045435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,1536,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,2048,0.6353466775682237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,128,0.0035146665241983202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,4096,0.011894221935007306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,5120,0.1824657784567939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,2048,0.046351999044418335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,128,0.009384888741705153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,16384,0.13248711162143284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,768,0.0068408888247278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,10240,0.1103973322444492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,2560,0.0698195563422309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,2560,0.012285333540704517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,512,0.0037395556767781577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,3584,0.011891555454995898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,64,0.16525333457522923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,3584,0.15367644362979466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,3072,0.01751377847459581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,1024,0.004058666527271271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,12288,0.4106728765699599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,128,0.0063866666621632045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,10240,0.022218666142887537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,1536,0.009209777745935652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,768,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,8192,0.13758222262064615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,5120,0.016526222229003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,1536,0.015249777171346875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,2048,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,10240,0.365739557478163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,256,0.003920888735188378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,3072,0.010981332924630908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,256,0.006105777704053455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,1024,0.40848623381720645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,6144,0.006271111054552927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,6144,0.06570577621459961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,2560,1.5318578084309895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,6144,0.10274577803081936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,3072,0.021306667062971327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,3072,0.016527111331621807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,1536,0.006069333189063602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,6144,0.04557866520351834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,2560,0.1729173395368788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,12288,0.286680883831448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,6144,0.008357333640257517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,10240,0.06977955500284831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,8192,0.3264595667521159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,32,0.0028595555987623003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,768,0.032398223876953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,65536,0.2125422159830729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,2048,0.005669333454635408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,1024,0.004271111140648524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,1024,0.025608888930744592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,1024,0.04562933246294657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,5120,0.008804444637563493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,128,0.0026622222115596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,3072,0.010527110762066312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,32,0.031140443351533677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,4096,0.12233511606852214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,256,0.003015999992688497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,3072,0.011233777635627322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,7168,0.027710222535663184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,3072,0.1272782219780816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,256,0.0032266666077905228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,3072,0.006866666591829724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,65536,2.5589377085367837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,16384,0.022650667362742957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,6144,0.026686223016844854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,5120,0.015059555570284525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,2560,0.12773244910769993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,16384,0.21318133672078451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,512,0.008090666598743862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,65536,5.077847798665364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,51200,0.18436888853708902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,1024,0.048102223210864596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,10240,0.3137253390418159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,7168,0.018279999494552612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,8192,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,3072,0.011692444483439127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,12288,0.46056710349188906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,3584,0.011705777711338468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,256,0.0226248883538776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,2048,0.059622221522861056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,7168,0.1679013305240207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,3072,0.0859911110666063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,12288,0.4856693479749892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,32,0.023670223024156358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,2048,0.006739555547634761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,1024,0.008032889001899296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,16384,0.049842665592829384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,65536,0.34799912240770126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,256,0.012039110892348819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,12288,0.3446177906460232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,7168,2.549744076199002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,64,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,32,0.0034337778472238113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,65536,1.1873306698269315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,16384,0.16920977168613008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,65536,0.25608711772494847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,3072,0.019488000207477145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,2560,0.023231110639042322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,1024,0.1721022261513604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,5120,0.012239110966523489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,1024,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,512,0.005581333405441708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,12288,0.0302817788388994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,4096,0.9533324771457248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,7168,0.3836969004737006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,1024,0.010341333018408881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,512,0.004864000197913912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,128,0.013600889179441663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,51200,0.9697928958468967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,4096,0.021120000216695998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,1024,0.018023111753993563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,5120,0.04895377821392483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,7168,0.014968888627158271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,256,0.004359999878538979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,65536,0.07155111100938585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,3584,0.011596444580290051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,8192,0.17902755737304688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,3584,0.00794044468137953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,5120,0.010115555591053432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,12288,0.0463066664006975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,5120,4.217976888020833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,1024,0.007864889171388414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,32,0.14187644587622747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,2048,0.009687110781669617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,12288,0.0221404449807273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,12288,0.06585244337717693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,256,0.02388266722361247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,16384,0.23144533899095324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,51200,21.799782647026912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,12288,0.08185333013534546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,5120,0.01814933286772834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,1536,0.017276444368892245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,12288,2.0158685048421225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,51200,1.685439109802246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,12288,0.1130711105134752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,768,0.008126222425036961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,8192,0.021980444590250652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,12288,0.03222222129503886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,3584,0.009699555734793345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,2560,0.13313599427541098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,10240,0.163665771484375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,128,0.006210666563775804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,768,0.005670222143332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,12288,0.30316087934705943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,5120,0.012949333422713809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,6144,0.014817777607176038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,10240,0.18939022223154703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,512,0.005777777896987067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,5120,0.04023733403947618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,768,0.010279110736317104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,16384,0.2377377880944146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,51200,0.6525866720411513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,1024,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,768,0.011139555937714048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,6144,0.1835555500454373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,5120,0.009750222166379293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,8192,0.028191109498341877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,10240,0.15230577521853977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,2560,0.056848890251583524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,256,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,1536,0.008744888835483128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,5120,0.018030222919252183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,256,0.0047333331571684945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,1024,0.004325333154863781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,8192,0.007894222107198503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,51200,2.1396363576253257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,7168,0.026147555973794725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,3584,0.020211555891566806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,7168,0.1771235598458184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,12288,0.05782577726576063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,3584,0.06452711423238118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,2560,0.027501333090994094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,10240,0.026641777820057336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,51200,1.6440248489379883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,8192,0.011784888803958893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,3072,0.20631021923489043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,5120,0.018610666195551556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,16384,0.019526221685939364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,10240,0.07496977514690824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,768,0.024264888630972967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,16384,0.3752604325612386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,8192,0.01684888866212633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,512,0.004292444636424382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,1024,0.02679288884003957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,1536,0.04162755608558655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,1024,0.006206222292449739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,64,0.003044444446762403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,51200,0.16548089186350504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,6144,0.0101724440852801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,128,0.021399110555648804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,1536,0.007735999921957652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,16384,0.014984000060293408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,128,0.11301510863833958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,8192,0.39954932530721027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,2560,0.24024176597595215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,512,0.021798221601380244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,256,0.013853333062595792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,4096,0.06557244724697538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,16384,0.03942844602796767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,16384,0.827822208404541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,10240,0.03532088796297709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,64,0.005625777774386936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,512,0.10202399889628093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,6144,0.012179555164443122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,4096,0.07281955745485094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,65536,2.417734146118164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,3584,0.27223287688361275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,10240,0.18049332830641004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,768,0.005872888697518243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,2048,0.187391996383667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,16384,0.13137688901689318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,512,0.00452888881166776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,512,0.0047120000753137796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,1536,0.006129777679840724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,5120,0.016241777274343703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,65536,0.5698764589097765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,3072,0.03882755504714118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,7168,0.01089422239197625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,12288,0.020630222227838304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,10240,0.39609511693318683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,1024,0.010115555591053432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,4096,0.010474666953086853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,2560,0.15241777896881104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,1536,0.021724444296624925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,2048,0.10968977875179714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,6144,0.017695999807781644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,4096,0.06078222062852648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,8192,0.03490488727887472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,3584,0.1362311045328776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,16384,0.09687466753853692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,768,0.006079999936951532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,128,0.002797333316670524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,3072,0.023973332511054143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,16384,0.02696711156103346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,3584,1.4022319581773546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,10240,0.008204444415039485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,2048,0.01772533357143402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,4096,0.07779288954204984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,6144,0.11828266249762641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,12288,0.2924871179792616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,2048,0.007864000068770515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,5120,0.008575110799736446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,3072,0.07226577732298109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,2048,1.5226017634073894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,256,0.006440889090299606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,128,0.008720888859695857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,1536,0.007099555598364935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,1536,0.08366400003433228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,6144,0.015040000279744467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,7168,0.01107377807299296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,51200,0.06200177801979912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,12288,0.04142133394877116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,5120,0.01774133410718706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,3584,0.026086222794320848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,16384,0.09649688667721218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,512,0.033901333808898926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,2560,0.015996444556448195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,128,0.016932444439993966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,256,0.008207999997668797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,512,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,1024,0.006079999936951532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,2560,0.0064106666379504735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,16384,0.020969778299331665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,7168,0.01404800017674764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,1024,0.00648266656531228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,2048,0.5243608686659071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,5120,0.08823466963238186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,10240,0.009199110998047723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,2048,0.011543999943468304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,51200,0.9388293160332574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,1024,0.00981866651111179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,6144,0.02214755614598592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,256,0.006172444257471297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,12288,0.619881788889567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,8192,0.009109333157539368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,4096,0.011013333168294696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,8192,2.9114373524983725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,1536,0.00700800038046307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,2048,0.017481777403089736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,10240,0.15698756111992732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,2560,0.009534222384293875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,3072,0.008409777449236976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,1024,0.2839084466298421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,51200,0.007909333540333642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,5120,0.10523644420835708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,16384,0.14164088832007513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,7168,0.0255031105544832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,1536,0.028909332222408716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,2560,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,3584,0.04044533438152737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,768,0.037596444288889565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,10240,0.06244799825880262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,12288,0.019303111566437613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,256,0.002847111059559716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,12288,0.396633783976237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,65536,0.1744542254341973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,512,0.006776888751321369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,32,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,256,0.048988444937600024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,2048,0.01316177762216992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,32,0.008188444707128737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,8192,1.1032550599839952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,7168,0.15246044264899358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,5120,0.07679200172424316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,10240,0.059193776713477246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,32,0.0029404444826973807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,2048,0.08907910850312974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,32,0.0026835555003749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,5120,0.037901331981023155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,1536,0.0047679998808436925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,6144,0.2398702303568522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,512,0.0033679999825027254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,128,0.012977778083748288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,8192,0.030124445756276447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,3584,0.008153777983453538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,12288,3.805487314860026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,2048,0.041053334871927895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,8192,0.09733866982989842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,12288,0.04299111167589823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,2048,0.03679733475049337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,7168,0.33303109804789227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,65536,0.06600888570149739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,16384,0.023039110832744177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,1536,0.0053084443012873335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,32,0.0026791110220882627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,51200,0.23695378833346895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,512,0.01237066669596566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,12288,0.06919022401173909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,3072,0.012089778151777057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,512,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,3072,0.1854728857676188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,256,0.010441777606805166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,3072,0.08220177888870239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,4096,0.03539022141032749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,8192,0.2727697690327962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,2560,0.030439111921522353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,4096,0.20045777161916098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,51200,0.3237066798739963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,256,0.005717333406209946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,2560,0.00777599960565567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,768,0.00815644446346495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,128,0.013616888887352414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,64,0.0029564443975687027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,256,0.14093511634402806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,51200,2.5928213331434464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,2560,0.17078844706217447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,16384,0.02244088881545597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,768,0.005452444569932089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,4096,0.14101510577731663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,768,0.06473510795169406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,7168,0.12128533257378472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,3584,0.05304355422655741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,5120,0.048159112532933555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,256,0.004602666530344221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,256,0.008316444853941599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,256,0.050960001018312245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,512,0.005984888722499211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,32,0.006572444405820634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,65536,0.2993146578470866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,256,0.00318755561278926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,12288,0.29161156548394096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,2048,0.008967111508051554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,1024,0.019144000278578866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,7168,0.349472893608941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,51200,0.3691546652052138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,51200,0.10460711187786526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,1536,0.016522667474216886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,5120,0.010978666444619497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,2048,0.007809778054555257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,3072,0.007435555259386699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,12288,0.020544000797801547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,1024,0.006722666737106111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,32,0.006624889042642381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,65536,0.09863377941979302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,16384,0.06929955879847209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,1536,0.004592888885074192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,32,0.015011555618709989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,1024,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,8192,0.2397057745191786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,4096,0.013247999880048962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,32,0.005021333280536863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,1024,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,2560,0.007681777907742395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,5120,0.013019555144839816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,256,0.004491555607981152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,5120,0.5636969142489964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,5120,0.012716444830099741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,8192,0.02055199940999349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,2560,0.0097884444726838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,64,0.0030382221771611106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,768,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,1536,0.0092257774538464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,64,0.003117333269781537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,2048,0.012516444755925072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,32,0.007662222617202335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,4096,0.008344000412358178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,16384,7.366385565863715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,512,0.0036862223512596558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,7168,0.0384808878103892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,512,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,2560,0.00832622249921163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,2560,0.05033777819739448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,2560,0.021581333544519212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,51200,0.5363608996073405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,128,0.005983110931184556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,6144,2.1276702880859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,256,0.014690667390823364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,12288,0.04042222102483114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,12288,0.016536000702116225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,64,0.0027591110103660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,128,0.005456888841258154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,12288,0.5357679790920681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,6144,0.014968888627158271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,2560,0.009360888765917884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,64,0.006525333142942852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,6144,0.006842666616042455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,3072,0.008678221868144141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,51200,0.3422977659437392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,2048,0.06434577703475952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,64,0.0025920000755124623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,12288,0.020097777247428894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,10240,0.32774845759073895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,128,0.004261333495378494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,2048,0.010527999864684211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,6144,0.03717777795261807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,2048,0.03703555464744568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,7168,0.07920355266994901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,4096,0.008567999634477828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,1024,0.02539644473128849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,3584,0.010214221974213919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,3584,0.10470222102271186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,1024,0.006835555450783835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,51200,3.055257797241211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,5120,0.01423911088042789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,3584,0.011667555405033959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,5120,0.5965920024447972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,6144,0.26929956012301975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,64,0.010339555641015371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,4096,0.009702222214804756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,12288,0.044043554200066455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,16384,0.1737555530336168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,1024,0.009353777600659264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,8192,0.006579555571079254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,65536,0.026260443859630164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,1024,0.28523021274142796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,1536,0.007146666447321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,128,1.2067271338568795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,512,0.004050666673315896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,512,0.004357333398527569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,16384,0.03098755412631565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,3072,0.02894222074084812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,10240,0.01369599997997284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,8192,0.05317155520121256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,10240,0.030983110268910725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,32,0.011136000355084738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,1024,0.031034668286641438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,768,0.017887999614079792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,128,0.017072889539930556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,128,0.0032177778581778207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,12288,0.0940542221069336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,16384,0.8076560232374402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,512,0.004104888687531154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,1024,0.006021333237489064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,8192,0.011742221812407175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,12288,0.21540710661146376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,1024,0.004836444639497333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,1536,0.0057706667317284485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,64,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,128,0.002687999978661537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,768,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,6144,0.017614222235149808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,2560,0.06378044684727986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,51200,0.060475554731157094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,6144,0.8841377894083658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,12288,0.7760062217712402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,10240,0.039162665605545044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,8192,0.8179297977023654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,2048,0.02104266650146908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,16384,0.30722933345370823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,2048,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,512,0.018989332848125033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,51200,2.274159961276584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,32,0.006036444256703059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,512,0.0107022225856781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,64,0.10621866914961074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,6144,0.30984266599019367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,16384,0.009278222090668148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,2560,0.0074133334888352295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,768,0.034613334470325045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,16384,1.648530642191569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,768,0.07070933447943793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,3584,0.01035911093155543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,51200,0.070814225408766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,768,0.38921154869927305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,12288,0.022477333744366963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,512,0.004915555732117759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,256,0.0034248888906505373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,8192,0.016688888271649677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,3584,0.008330666356616551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,1024,0.03171644608179728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,3584,0.02461333407296075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,7168,0.39058576689826113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,512,0.003933333274390963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,512,0.006173333360089197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,3072,0.07108355230755277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,10240,0.22502843538920084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,1024,0.008267555799749162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,1536,0.03388266762097677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,16384,0.20361867215898302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,1536,0.006311111152172089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,2560,0.013593778014183044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,4096,0.008784888519181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,4096,0.007754666937722101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,10240,0.03647822141647339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,12288,0.033859554264280535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,256,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,2048,0.007666666474607255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,12288,0.01399466726515028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,12288,0.03239288926124573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,16384,0.955259535047743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,2048,0.013574221895800697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,65536,0.2887626753913032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,4096,0.4285031159718831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,768,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,256,0.008246222303973304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,3072,0.06426755587259929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,2560,0.012190221912331052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,8192,0.031437334087159895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,256,0.08664977550506592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,12288,0.041906668080223926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,8192,0.5445306566026475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,1024,0.006816888848940532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,512,0.005333333379692501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,8192,0.0288702216413286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,64,0.0033200000309281876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,51200,0.059586664040883384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,2048,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,3072,0.05794044335683187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,5120,0.10611555311414932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,12288,0.02410488824049632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,10240,1.13500001695421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,6144,0.028424001402325098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,5120,0.29527823130289715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,4096,0.014352000421947904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,256,0.0111191107167138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,16384,0.08198666572570801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,7168,0.03229066729545593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,1536,0.012244444754388599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,768,0.052419556511773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,1024,0.08308621909883288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,3584,0.007784888976150089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,3072,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,128,0.05227111114395989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,2560,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,2560,0.009218666288587783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,2048,0.020032000210550096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,768,0.030771556827757094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,3072,0.006477333191368315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,2560,0.017875555488798354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,768,0.003917333152559069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,3072,0.03345155715942383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,12288,0.6743546591864692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,3072,0.08972533543904622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,2048,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,256,0.0063484443558586975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,768,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,12288,5.24510023328993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,6144,0.346921788321601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,2048,0.01850666602452596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,128,0.006187555690606435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,512,0.02272266646226247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,2560,0.1775395605299208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,5120,0.008775110873911116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,8192,0.029874665869606867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,5120,0.023394667439990576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,16384,0.4249448776245117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,51200,0.060384889443715416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,64,0.02197066611713833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,16384,0.14475733704037136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,2560,0.014759999182489185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,128,0.004295999805132548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,2560,0.02292888859907786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,3584,0.02055199940999349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,8192,0.013349333571063148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,2048,0.0053591111467944244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,768,0.003899555653333664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,6144,0.09453866879145305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,64,0.024869332710901897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,1536,0.004680888934267892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,8192,0.015837333268589444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,8192,0.20107021596696642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,64,0.03191288974550035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,2048,0.038520889149771795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,10240,0.17315110895368788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,65536,0.21082488695780435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,8192,0.019732443822754752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,51200,0.020548444655206468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,6144,1.1431271235148113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,256,0.0053839998112784485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,5120,0.25733423233032227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,10240,0.03827733463711209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,4096,0.016944888565275405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,8192,0.08381511105431451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,16384,0.024143111374643113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,2560,0.008435555630260045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,51200,0.45238934622870547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,7168,0.08920266893174912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,5120,0.01849244369400872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,10240,0.05432977941301134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,3584,0.06339733468161689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,7168,0.12854311201307508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,3072,0.12447288301255967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,65536,15.490697224934896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,768,0.016502221425374348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,5120,0.2431404325697157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,6144,0.029024001624849107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,10240,0.4022391107347276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,3072,0.018295110927687753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,4096,0.09928177462683783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,32,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,3072,0.01109955542617374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,6144,0.013587555951542325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,5120,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,64,0.008200888832410177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,2560,0.010928000013033548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,8192,0.03537866804334853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,768,0.006164444403515921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,3584,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,512,0.07250222232606676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,256,0.14691555500030518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,4096,0.014905777242448596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,1536,0.007752888732486301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,64,0.005327111317051782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,128,0.02769333289729224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,128,0.0027546667390399506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,768,0.007675555845101674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,2560,0.01886399918132358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,16384,0.027874665127860174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,65536,0.06572800212436251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,768,0.025359110699759588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,2560,0.17456977897220186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,256,0.005554666535721884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,2560,0.034274667501449585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,6144,0.05117688907517327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,512,0.003965333518054751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,128,0.01240355521440506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,16384,9.142185635036892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,2048,0.07106755839453803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,51200,0.1733671161863539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,2560,1.8442889319525824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,3584,0.07930844359927706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,10240,0.17561777432759604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,8192,0.04811288913091024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,4096,0.03128177920977274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,1024,0.007050666544172499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,3584,0.007325333025720384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,1024,0.016905777984195285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,7168,0.016333333320087857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,51200,2.4036142561170792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,6144,0.013992889059914483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,1536,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,10240,0.012344000240166983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,3584,0.024460444847742718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,3072,0.1650924417707655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,1024,0.013677332964208392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,3072,0.009635555247465769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,4096,0.04831466741032071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,3584,0.007910221815109253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,3072,0.025801777839660645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,8192,0.018601778480741713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,4096,0.04423822296990288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,12288,2.842816882663303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,12288,0.02552977701028188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,4096,0.047994666629367404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,8192,0.12193244033389622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,2048,0.03904089000489976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,1536,0.02537333303027683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,10240,0.03100088900989956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,10240,0.11477422714233398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,8192,0.29111022419399685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,2048,0.3176151116689046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,8192,0.007495111061467065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,2048,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,32,0.00795999997191959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,128,0.01254755589697096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,1536,0.006412444429265128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,3072,0.046210666497548424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,4096,0.026944888962639704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,65536,1.205201784769694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,1536,0.16944621668921578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,1024,0.02146311104297638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,65536,0.5507520039876302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,1536,0.025605332520272996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,768,0.0064488889442549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,4096,0.0063075555695427795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,8192,0.017266665895779926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,51200,0.12305688858032227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,1536,0.1007573339674208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,768,0.01889333294497596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,12288,0.09408622317843968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,1024,0.01093155559566286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,8192,0.05383111039797465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,64,0.0380311111609141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,5120,0.022239110536045496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,512,0.007872000336647034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,12288,0.11044355233510335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,16384,0.0070053330726093715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,64,0.011420444481902652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,3584,0.00830488900343577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,7168,0.04418311185306973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,12288,0.12652266025543213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,51200,1.5181893242730036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,12288,0.04746133420202467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,2048,0.016187555260128446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,2560,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,768,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,128,0.03681688838534885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,4096,0.011979555918110741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,2560,0.8171351220872668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,5120,0.07400711377461751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,256,0.013045333325862885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,2048,0.01260444439119763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,64,0.0028293333533737394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,1536,0.00444177786509196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,2560,0.057251552740732826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,7168,0.6043253474765354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,128,0.003882666842805015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,128,0.005054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,10240,0.2227271132998996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,8192,0.006702222343948152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,2048,0.008396444221337637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,2048,0.012303111453851065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,1024,0.021734222769737244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,1536,0.009736888938479954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,12288,4.33646477593316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,7168,0.02085599965519375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,10240,0.1666942172580295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,65536,0.6904133160909017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,256,0.010821333361996545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,3584,0.04055644406212701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,32,0.045019557078679405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,3584,0.03887733485963609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,3072,0.007439999944633908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,12288,0.060419552856021456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,65536,17.689405653211804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,128,0.0047111109726958806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,16384,1.360372437371148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,128,0.012629333469602795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,512,0.004705777598751916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,4096,0.04153866569201151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,128,0.0041404445138242514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,1024,0.01556711064444648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,128,0.15096799532572427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,256,0.01055022246307797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,768,0.032398223876953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,512,0.005975111077229182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,16384,0.05729599793752035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,5120,0.009350222018029954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,4096,0.08728088935216267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,8192,4.971337636311849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,64,0.010112000008424124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,1024,0.008933333059151968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,768,0.004376000000370873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,256,0.0033368888414568375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,5120,0.06564533048205905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,8192,0.049365331729253135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,10240,0.03487022386656867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,10240,0.17799999978807238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,3072,0.008647999829716152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,6144,0.603424866994222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,2560,0.022043555974960327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,256,0.007415999968846639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,6144,0.010090666512648264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,16384,1.2370239893595378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,12288,0.3095093303256565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,1536,0.008606221940782335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,2560,0.24040354622734916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,1536,0.016566221912701923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,16384,0.2698791027069092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,51200,26.87752956814236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,2048,0.01073155552148819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,3072,0.06863911284340753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,51200,4.2702818976508246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,12288,0.49454667833116317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,3584,0.08379199769761826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,1024,0.015976889265908133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,1536,0.23073244094848633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,12288,2.4788345760769315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,512,0.04362844427426656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,6144,0.21734400590260824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,512,0.006076444354322221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,2560,0.27030576599968803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,4096,0.0064791109826829694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,2048,0.005506666584147348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,6144,0.01352533366945055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,7168,0.01610311037964291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,6144,0.03747111227777269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,3072,0.2382462289598253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,3584,0.01993333299954732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,768,0.006963555183675554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,5120,0.01278222186697854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,128,0.004872888740566042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,6144,0.0517759985393948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,128,0.0027955555253558685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,4096,1.820150163438585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,3584,0.008910222185982598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,1536,0.006561777657932705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,3072,0.21047555075751412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,64,0.0058542220956749385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,16384,0.8074391153123645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,1024,0.03687199950218201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,8192,0.20344444115956625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,256,0.006543111056089401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,768,0.007544000115659501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,65536,1.3541235393948023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,128,0.01291911138428582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,6144,0.017344888713624742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,2048,0.009187555147541894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,51200,0.5618684556749132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,7168,0.05893599987030029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,3072,0.048375109831492104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,16384,0.44002310434977215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,1024,0.016951999730534024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,16384,2.148136986626519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,2560,0.010481778118345471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,64,0.0032702222880389956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,10240,0.10369866424136692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,768,0.004952888935804367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,32,0.0032986667421129015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,16384,0.5307262208726671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,1024,0.0399600002500746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,768,0.010749333434634738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,7168,0.27202754550509983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,2560,0.02048711147573259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,5120,0.014314666390419006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,3584,0.031522668070263334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,3584,0.10778044329749213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,64,0.003310222178697586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,10240,0.08360799815919663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,10240,0.033590220742755465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,16384,0.059902224275800914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,6144,0.010644444160991244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,64,0.004453333301676644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,10240,0.19676444265577528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,6144,0.8212800025939941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,51200,0.20768445067935518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,65536,2.2972524431016708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,3072,0.010517333116796283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,128,0.010619555910428366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,256,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,6144,0.11558932728237575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,7168,0.8264400164286295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,1536,0.04444977641105652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,32,0.014387556248241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,256,0.008894222478071848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,51200,0.5526151127285427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,32,0.019339554839664035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,6144,0.027718222803539697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,3072,0.02481777800454034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,32,0.002999111182159848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,256,0.009533333281675974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,7168,0.14713154898749456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,768,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,16384,0.07668355438444349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,10240,0.9204275343153211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,3584,0.009711999860074785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,3072,0.01166488892502255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,8192,0.022062222162882488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,512,0.176454226175944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,16384,0.10665511422687107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,8192,0.011839110818174152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,5120,0.09161688884099324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,6144,0.0564604467815823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,5120,2.0088488260904946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,10240,0.27182398902045357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,128,0.009453333086437648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,1536,0.021873777111371357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,1536,0.23805954721238878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,512,0.027064000566800434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,768,0.049591110812293164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,8192,2.247720930311415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,2560,0.2765271133846707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,512,0.045885334412256874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,6144,0.010812444819344414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,65536,0.08419555425643921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,128,0.00300177786913183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,4096,0.041144887606302895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,2560,0.06587911314434476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,7168,0.22586666213141549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,10240,0.015965332587560017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,10240,0.012871111432711283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,5120,0.008887111312813228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,4096,0.00870488915178511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,2560,0.022476444641749065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,8192,0.12090222040812175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,10240,0.07838666439056396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,2048,0.1253946622212728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,3584,0.032863110303878784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,512,0.0046844445168972015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,3072,0.007401777638329401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,8192,0.04064977831310696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,3584,0.026176000634829204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,12288,0.2770284546746148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,3072,0.02567022211021847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,256,0.0058764442801475525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,128,0.0047333331571684945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,3584,0.018352000249756705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,12288,0.08055111434724596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,256,0.00833511104186376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,6144,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,256,0.00814222213294771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,1536,0.00601244428091579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,2048,0.05194044444296095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,4096,0.028093334701326158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,768,0.005614222337802251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,512,0.005218666460778978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,8192,0.6892053286234537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,51200,0.33336978488498265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,10240,0.5526880158318414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,7168,0.01295111080010732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,64,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,256,0.006677333265542984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,512,0.018064889642927382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,256,0.006232888748248418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,32,0.05424000157250298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,3584,0.028918223248587713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,65536,0.05617600017123752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,8192,0.07154666715198092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,768,0.007855110698276097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,8192,0.06111022498872545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,3584,0.03065955638885498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,51200,0.26643821928236217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,32,0.0027964444210131965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,128,0.004855999930037392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,2560,0.04526577724350823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,3072,0.01851466629240248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2048,32,0.005009777843952179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,512,0.004559111264016894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,2560,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,8192,0.011620444556077322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,1024,0.011553777588738335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,1024,0.11276088820563422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,1536,0.004948444250557157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,3584,0.010260444548394945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,256,0.117576003074646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,128,0.004481777962711123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,1536,0.007076444725195567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,6144,0.7669910854763455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,256,0.0035982223020659555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,2048,0.0689893364906311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,1536,0.01129244433508979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,5120,0.06359199682871501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,5120,0.0916995538605584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,2560,0.0159582214223014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,65536,1.4420479668511286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,768,0.005764444255166584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,64,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,2560,0.009890666438473595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,512,0.027375110321574744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,8192,2.5936525132921004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,4096,0.25858844651116264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,768,0.026320889592170715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,64,0.0052595556610160405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,512,0.0053342220683892565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,5120,0.10421688689125909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,768,0.011634666886594562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,2560,0.011202666494581433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,2048,0.08832889133029515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,2048,0.3907795482211643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,128,0.004495999879307217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,256,0.021334222621387903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,2048,0.011306666665607028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,65536,0.15449866983625624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,1024,0.056031107902526855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,4096,0.008794666992293464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,3584,0.09384800328148736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,1024,0.0058115555180443665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,3072,0.10278044144312541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,256,0.004135999828577042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,6144,0.013351110948456658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,7168,0.02001422146956126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,5120,0.038888888226615057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,1536,0.014154665999942355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,2048,0.043466667334238686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,32,0.06056266360812717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,1536,0.016145777371194627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,2048,0.00675644435816341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,5120,0.014214222629865011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,4096,0.01347644461525811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,10240,0.11637066470252143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,7168,0.18444621562957764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,12288,0.05209155546294319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,2048,0.0075733330514695905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,3072,0.5036151144239638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,768,0.003748444633351432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,128,0.0042408886882993905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,5120,0.3248319890764025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,1024,0.006160888820886612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,128,0.0032133333798911837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,1024,0.11340178383721246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,128,0.009438222481144799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,3584,0.00906399968597624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,768,0.007976888782448238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,3584,0.01218844453493754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,3072,0.015346666177113851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,1536,0.0074577778577804565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,64,0.005748444547255834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,16384,0.24814134173923066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,10240,0.1381431155734592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,1536,0.008187555604510838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,256,0.006582222051090664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,128,0.12320889366997613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,5120,26.161300659179688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,3072,0.026292444931136236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,4096,0.023346667488416035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,3584,0.008068444828192392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,2560,0.02618933386272854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,8192,0.0188391109307607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,64,0.010823110739390055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,6144,0.016082665986484952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,4096,0.10064178042941624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,1536,2.6175734202067056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,7168,0.014641778336630927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,4096,0.057816889550950795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,128,0.00852888905339771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,3072,0.01568711135122511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,32,0.005537777725193236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,6144,0.01144088887506061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,65536,0.6360524495442709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,32,0.0035946667194366455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,768,0.01127022256453832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,1024,0.07906399832831489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,65536,0.08583199977874756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,65536,0.01186044431395001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,1536,0.3084782229529487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,12288,0.08022222254011366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,4096,0.06297866503397624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,1024,0.09708799918492635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,128,0.007742221984598372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,16384,0.5652053091261122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,51200,0.024674667252434626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,12288,0.019864888654814828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,2560,0.5014630953470866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,8192,0.035346666971842446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,51200,0.06360089116626315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,4096,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,1536,0.02593333356910282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,3072,0.01290933373901579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,256,0.0414631101820204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,12288,0.17495289113786486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,6144,0.0258240004380544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,512,0.05748889181349012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,32,0.011320888996124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,2048,0.009688888986905416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,1536,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,3584,0.0714977780977885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,4096,0.14608089129130045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,10240,0.1629884507921007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,128,0.0027715555495686005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,8192,0.03644711110326979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,1024,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,1536,0.018311111463440787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,4096,2.48507202996148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,10240,0.08957511186599731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,2560,0.7993546591864692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,768,0.008056888977686564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,4096,0.02862488892343309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,3072,0.00756444450881746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,1024,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,8192,0.012904889053768583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,10240,0.016574222180578444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,3072,0.019501333435376484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,1024,0.008477333519193862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,512,0.028504000769721136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,768,0.006034666465388403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,2560,0.044663998815748424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,64,0.0029395555870400537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,51200,0.36251555548773867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,6144,0.03717422154214647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,4096,0.41379555066426593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,6144,0.013223110801643796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,51200,0.0933084421687656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,12288,0.013955555028385587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,2560,0.008787555827034844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,5120,0.3754817909664578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,5120,0.251618676715427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,16384,0.02850666642189026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,3584,0.0101742222905159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,51200,0.42224444283379453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,65536,0.4545920160081651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,8192,0.011961778004964193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,2560,0.007727111379305522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,12288,0.018944889307022095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,8192,0.0214035560687383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,64,0.02865333358446757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,768,0.011687111523416309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,768,0.0038933331767717996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,65536,0.23728267351786295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,8192,0.024566221568319533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,1024,0.008470222353935242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,51200,0.1892577807108561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,512,0.0035297779573334586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,6144,0.07785599761539035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,2048,0.010755555497275459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,10240,0.03496355480617947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,1024,0.00794400026400884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,6144,0.025791999366548326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,768,0.06824622551600139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,768,0.006056889063782162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,12288,0.16132267316182455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,128,0.009881777895821465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,2560,0.10581066873338486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,32,0.005865777946180767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,5120,0.2829751173655192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,3584,0.015061333775520325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,2048,0.02030666669209798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,10240,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,7168,0.019966221517986722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,2048,0.04225688841607836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,2560,0.030135111676322088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,64,0.003059555465976397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,51200,0.058131555716196694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,7168,0.023376888699001733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,4096,0.009359111388524374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,64,0.0029502221279674103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,7168,0.06298310889138116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,7168,0.35451022783915204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,12288,7.417667812771267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,64,0.004218666503826777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,5120,0.02829688787460327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,12288,0.013550221920013428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,128,0.01587288909488254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,1024,0.10712977912690903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,4096,0.017966222431924608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,32,0.010248888697889116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,1536,0.009139555195967356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,256,0.01780533293883006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,512,0.02479822271400028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,16384,0.5721928808424208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,2048,0.01625155574745602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,4096,0.03296355406443278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,51200,0.07770133018493652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,128,0.005662222289376789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,32,0.009141333401203156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,8192,0.015747555428081088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,8192,3.9752650790744357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,1536,0.037250667810440063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,3584,0.028096887800428603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,12288,0.0421688887808058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,3584,0.012373333175977072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,7168,0.10214933421876694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,4096,0.012265777422322167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,2048,0.018938667244381376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,128,0.006096000058783426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,8192,0.15825866328345403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,1024,0.0041511112617121804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,65536,0.0742151141166687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,4096,0.054810666375690036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,3072,0.049794667296939425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,7168,0.028963555892308552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,3584,0.04741866721047295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,32,0.002564444517095884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,6144,0.00629688882165485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,256,0.017800889081425138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,5120,0.029993779129452173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,32,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,51200,1.4179342057969835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,65536,0.16023377577463785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,2560,0.019965334071053397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,128,0.0046160001721647055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,2048,0.008392888638708327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,128,0.005903111149867375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,8192,0.035565333233939275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,6144,0.010443555812040964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,1536,0.006324444380071428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,512,0.004586666822433472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,64,0.013454222016864352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,3072,0.20690666304694283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,6144,0.013916444447305469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,768,0.11733867062462701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,2048,0.01901333365175459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,6144,0.029191109869215224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,1536,0.0061919999619325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,51200,30.956281873914932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,1024,0.035595556100209556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,3584,0.04885511265860664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,256,0.006315555423498154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,1024,0.00850577818022834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,6144,0.09216266870498657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,12288,3.2981715732150607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,12288,0.2240755558013916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,51200,0.7091777589586047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,256,0.1617386606004503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,1024,0.0064399999876817065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,64,0.002623999905255106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,64,0.03687377770741781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,8192,0.029640889830059473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,7168,0.34387021594577366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,65536,1.7026809056599934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,6144,0.01882488860024346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,7168,0.1049297783109877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,3072,0.00777599960565567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,5120,0.25398132536146373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,12288,0.01812622282240126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,16384,0.11547999911838108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,2560,0.009518221848540837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,8192,0.12546133995056152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,1024,0.02565333412753211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,128,0.00453599997692638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,65536,1.7611697514851887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,12288,0.06797333558400472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,3584,0.0199297782447603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,51200,0.706241766611735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,8192,0.19392889075809053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,65536,0.08253600200017293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,6144,0.01165422217713462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,8192,0.014582221706708273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,12288,0.015142222245534262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,256,0.007071110937330458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,65536,0.27883021036783856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,2560,0.1425386667251587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,16384,0.1406373315387302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,12288,0.07955377631717257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,2560,0.0075048887067370946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,768,0.15663288699256048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,128,0.008110221889283922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,8192,0.04687111245261299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,2048,0.00648266656531228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,2560,0.1312266720665826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,768,0.003666666646798452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,5120,0.015218666858143277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,64,0.009754666851626502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,4096,0.013213333156373767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,51200,0.3198755582173665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,5120,0.015079110860824585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,16384,0.04610222246911791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,2560,0.23470044136047363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,1536,0.010695999695195092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,5120,0.012597333225939008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,128,0.0299288895395067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,7168,0.23505332734849718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,1024,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,51200,0.219869335492452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,1024,0.04071911176045736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,7168,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,2560,0.00979288915793101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,10240,0.1491831143697103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,12288,0.01622488929165734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,1024,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,12288,0.09089422225952148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,512,0.0033457777980301115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,65536,4.010986751980252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,6144,0.03753688931465149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,512,0.004046222401989831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,3072,0.02811288833618164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,128,0.005984000033802456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,51200,0.5188310941060384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,7168,0.2926222218407525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,8192,0.15611999564700657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,4096,0.010651555326249866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,2560,0.011047999891969891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,3072,0.02089244458410475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,65536,8.453646341959635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,512,0.004836444639497333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,128,0.017469333277808297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,10240,0.01612622208065457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,12288,0.15027111106448704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,1536,0.031587554348839655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,1024,0.006624889042642381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,3584,0.01178311142656538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,2560,0.6020888752407497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,12288,0.0350471105840471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,2560,0.017441777719391715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,3584,0.007305777735180325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,1024,0.021719111336602107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,64,0.005668444352017508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,128,0.03209688928392198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,16384,3.0872444576687283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,10240,0.2883235613505046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,256,0.006141333530346553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,8192,0.050352000527911715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,4096,0.2594124476114909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,512,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,6144,0.014727999766667684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,3072,0.01464355488618215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,5120,0.014222222897741528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,768,0.03937866621547275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,512,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,2560,0.09919644726647271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,768,0.0443173348903656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,256,0.0041617775956789655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,256,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,6144,0.5848222308688694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,3072,0.7995520167880588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,3584,0.019529778096410964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,7168,0.18669422467549643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,2560,0.006346666564544042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,1536,0.011202666494581433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,768,0.006634666687912411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,768,0.08280266655815972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,256,0.004902222090297275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,16384,0.6758009062872993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,32,0.0029813332690132987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,512,0.02626666592227088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,256,0.003302222324742211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,12288,0.011324444578753578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,7168,0.014331556028789945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,10240,0.24198844697740343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,128,0.007128888534175024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,16384,0.0248364441924625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,4096,0.0820853312810262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,512,0.003738666574160258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,64,0.026718222432666357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,65536,0.07690755526224773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,8192,0.07041688760121663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,768,0.0036319999231232535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,512,0.03964799973699782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,1024,0.010636444720957013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,768,0.037455999188952975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,5120,0.09590222438176473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,512,0.026554667287402686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,64,0.0052728888889153796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,256,0.009024000002278222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,1536,0.10466755761040582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,2048,0.0681333343187968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,128,0.004607111215591431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,128,0.0035342222286595237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,10240,0.012812444733248817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,12288,0.1538728872934977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,12288,0.1192728943294949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,6144,0.06875199741787381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,768,0.008496888809733922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,256,0.010252444280518426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,3584,0.010880889164076911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,12288,0.07683555285135905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,512,0.017157332764731515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,65536,8.129570855034723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,51200,0.7207733260260688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,1024,0.05795377492904663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,512,0.0041413332025210066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,128,0.009882666998439366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,3072,0.01701866587003072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,6144,0.16710400581359863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,2560,0.07063555717468262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,65536,0.20472799407111275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,3584,0.011191111471917896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,512,0.013600000076823764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,1536,0.007799111306667328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,128,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,5120,0.15282489193810359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,3584,0.2772257857852512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,7168,0.01664622293578254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,32,0.003995555556482739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,1024,0.006017777654859755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,3072,0.013891556196742587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,32,0.014692443940374585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,3072,0.009383111364311641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,512,0.008088000118732452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,2048,0.008021333151393467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,3072,0.01687022215790219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,6144,0.014872888724009195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,16384,0.052561776505576245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,64,0.002529777793420686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,5120,0.016495111915800307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,10240,0.33067554897732204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,16384,0.030365334616767034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,32,0.005193777796294955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,768,0.0046844445168972015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,3584,0.12080978022681342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,65536,0.09328089157740276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,256,0.04053066505326165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,2048,0.01573333309756385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,1024,0.006170666466156642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,51200,0.08406133121914333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,1536,0.013762666947311826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,32,0.003044444446762403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,16384,0.15511288907792833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,1024,0.09173778030607437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,7168,0.010996444357766045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,5120,0.012972444295883179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,3584,0.03769955701298184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,6144,0.19687022103203666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,128,0.026783999469545152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,512,0.007594666547245449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,3072,0.020776000287797716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,65536,183.64720323350696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,1024,0.04058666692839728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,2048,0.010873777998818291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,64,0.002512000087234709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,1536,0.0357528891828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,1536,0.017177777157889474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,12288,0.03155022197299533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,4096,0.9546542697482638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,1536,0.011365333365069495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,1536,0.019504888190163504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,6144,0.013615999784734515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,64,0.1325440009435018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,8192,0.06694044669469197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,1024,0.005960888746711943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,7168,0.04831466741032071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,51200,4.479520161946614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,768,0.01549600064754486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,12288,0.3608711030748155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,65536,0.1250337759653727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,51200,0.1966195503870646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,2048,0.05541955431302389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,8192,0.030590222941504583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,65536,0.4184346728854709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,16384,0.006189333481921091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,8192,0.0454693337281545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,8192,0.011735111474990845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,8192,0.0638071099917094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,65536,0.06780533658133613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,128,0.008016000191370646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,1536,0.04643288917011685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,256,0.015826667348543804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,65536,0.07804889149136014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,4096,0.010980444649855295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,16384,2.6620000203450522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,512,0.00432977784011099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,16384,0.0718053314420912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,8192,0.32687287860446507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,1024,0.038563556141323514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,2048,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,256,0.008064889245563084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,51200,0.41528089841206867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,2560,0.036972443262736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,4096,0.013473778135246701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,512,0.013296888934241401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,512,0.00463733325401942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,3072,0.01754577789041731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,2048,0.007227555745177799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,32,0.002575111058023241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,5120,1.6279332902696397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,8192,0.014178666803571912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,5120,0.01661066710948944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,768,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,16384,1.0162942674424913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,512,0.006980444822046492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,10240,0.01881866653760274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,1024,0.01667555504375034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,10240,0.04574311110708448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,6144,0.015955555770132277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,10240,0.042167110575569995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,4096,0.016375111209021676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,64,0.0033039999090962936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,16384,0.051408890220854014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,7168,0.014965333872371249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,51200,0.157496001985338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,3072,0.006776888751321369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,5120,0.048231111632453076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,65536,21.570220947265625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,1024,0.0090515555606948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,1536,0.7509102291531033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,51200,2.234936820136176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,8192,0.11239377657572429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,12288,0.3617902331882053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,16384,0.038691557115978666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,6144,0.01641688909795549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,2048,0.007912000020345053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,5120,0.030438221163219873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,10240,0.06677777899636163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,512,0.003636444608370463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,3072,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,128,0.009305777649084726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,16384,6.371230231391059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,51200,1.017807960510254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,6144,0.03418666786617703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,3072,0.023951111568344965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,65536,0.08004444175296359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,2048,0.03492888808250427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,65536,0.2445920043521457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,3072,0.03386577632692125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,64,0.003104000041882197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,51200,0.34940534167819554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,32,0.00945955514907837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,10240,0.04781422350141737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,51200,0.05564622084299723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,1024,0.007380444142553542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,1536,0.04048711061477661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,65536,0.35373332765367294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,1536,0.006311111152172089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,8192,2.093936072455512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,5120,0.19949155383639863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,4096,0.02104444470670488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,1024,0.028099556763966877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,1536,0.03279022375742594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,1536,0.03955111238691542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,8192,0.032235556178622775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,6144,0.32634221182929146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,3072,0.19550844033559164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,51200,0.19735556178622773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,64,0.002558222247494592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,64,0.005398222141795688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,2560,0.0064035554726918536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,768,0.014128888646761576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,10240,0.026896889011065166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,8192,0.20903289318084717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,6144,0.020523554748959012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,2048,0.005980444451173146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,6144,0.026632888449562922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,1024,0.01940266622437371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,2048,0.00851288851764467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,512,0.0053688887920644544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,5120,0.012783110969596438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,32,0.0030133333057165146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,4096,0.22953422864278158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,8192,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,128,0.03565066721704271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,768,0.041929778125551015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,3584,0.0555937753783332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,65536,0.022783999641736347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,12288,0.0191600008143319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,12288,0.4139022297329373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,2048,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,1536,0.02181688944498698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,16384,0.012225777738624148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,3584,0.00978488889005449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,65536,0.0397093329164717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,2560,2.236612531873915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,2560,0.09623999728096856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,51200,0.009132444030708736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,51200,4.166634877522786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,2560,0.2223270999060737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,1024,0.04517244299252828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,512,0.021937777598698933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,7168,0.007374222079912822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,6144,0.021701332595613267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,4096,0.008013333711359236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,16384,0.015857777661747403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,1536,0.524328867594401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,10240,0.026375111606385972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,512,0.01664266652531094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,128,0.0047191112405723995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,51200,0.14844622876909044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,2560,0.20497600237528482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,2048,0.025794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,5120,0.1157182190153334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,65536,0.12486578358544244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,12288,0.6854764620463053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,12288,0.43886754247877335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,512,0.010509333676762052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,2560,0.017445334129863314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,4096,0.015103111664454142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,7168,0.09855822059843276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,2048,0.03338311115900675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,5120,0.0062773335311147906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,2560,0.010873777998818291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,7168,0.23622133996751574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,3072,0.0074275558193524676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,2048,0.014744889405038623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,10240,0.01623288955953386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,4096,0.05714133050706652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,128,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,128,0.010077333284748925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,768,0.03017688790957133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,8192,0.4021466573079427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,10240,0.07973421944512261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,5120,0.09262666437360977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,3584,0.27092356152004665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,32,0.008700444466537898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,3072,1.007881800333659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,7168,0.0166284441947937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,3584,0.011137777732478248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,768,0.4505031373765733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,128,0.003916444463862313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,3584,0.17518222332000732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,65536,1.5711857477823894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,8192,0.032447109619776406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,1024,0.012797333300113678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,2560,0.00563733321097162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,12288,5.649094899495442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,768,0.03881511092185974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,10240,0.09509422381718953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,768,0.49693775177001953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,256,0.02255288925435808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,3072,0.013056889176368713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,65536,0.08274844619962904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,51200,0.17407822608947754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,3584,0.020967110991477966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,1536,0.01202666676706738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,16384,0.05271466573079427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,12288,0.18425333499908447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,1024,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,512,0.0041999999019834726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,2560,0.19371555911170113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,12288,0.04068533248371548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,65536,0.07004355722003512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,1024,0.042691555288102895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,65536,0.07182577583524916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,32,0.07791466845406426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,64,0.00499288903342353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,6144,0.012183999849690331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,12288,0.028159111738204956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,1024,0.007680888805124495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,512,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,512,0.0033813334173626373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,2048,0.008404444489214156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,10240,0.016851555969980028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,768,0.0036924444139003754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,65536,0.47544087304009336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,3072,0.5310062302483453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,2560,0.057180444399515785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,64,0.00702400008837382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,2560,0.012531555361217923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,512,0.007735110819339752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,64,0.005161777966552311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,6144,0.01906133360332913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,256,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,1536,0.033442666133244835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,1536,0.008552889029184977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,2048,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,3584,0.022260444031821355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,2048,0.00960177762640847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,1536,0.004924444274769889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,16384,0.043288889858457774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,16384,0.20012266106075713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,7168,0.025123554799291823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,51200,2.720081753200955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,12288,2.160909228854709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,12288,0.014679999815093147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,64,0.0032435556252797446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,3584,0.011888888974984487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,4096,0.01700088878472646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,512,0.003666666646798452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,10240,0.012627555264366997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,51200,0.16793421904246011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,65536,0.06869244575500488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,2560,0.18116177452935112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,7168,1.2355208926730685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,768,0.024697777297761705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,5120,0.05865066581302219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,65536,0.4348746670616998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,768,0.010341333018408881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,128,0.005068444543414646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,256,0.005642666584915585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,256,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,6144,0.006631111105283101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,5120,0.17945510811275908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,3584,0.012119111087587146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,4096,0.031567109955681696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,8192,0.07336444324917264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,10240,0.3646995491451687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,51200,0.9001760482788086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,3584,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,51200,0.13720444838205972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,65536,0.13837599754333496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,10240,0.013579555683665805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,7168,0.03515999846988254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,1024,0.018206222189797294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,8192,3.786810557047526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,1536,0.0758240024248759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,4096,0.1755919986300998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,2560,0.01611111064751943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,3584,0.15016710758209229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,4096,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,4096,0.08112088839213054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,768,0.0099973330895106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,128,0.005424889011515512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,4096,0.013196444345845116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,64,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,128,0.013835555149449242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,1024,0.003979555434650845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,51200,0.07023111316892836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,51200,0.5122710863749186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,16384,0.072708441151513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,1536,0.23216711150275338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,1536,8.51864284939236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,10240,1.1373795403374565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,2048,0.060025778081681996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,2048,0.00941777808798684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,768,0.004765333400832282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,2560,0.0330079992612203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,51200,12.851852416992188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,5120,0.014239999983045789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,3072,0.02113688819938236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,16384,8.129347059461805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,256,0.04330577784114414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,7168,0.5140213436550564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,4096,0.03305155701107449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,10240,0.08670755227406819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,256,0.03124533428086175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,1536,0.024028443627887305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,10240,0.018279999494552612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,2048,0.0404097768995497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,3072,0.0839795536465115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,2560,0.17025955518086752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,6144,0.09723555379443699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,10240,0.014141332772043018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,12288,0.9760622448391385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,5120,0.03613866700066461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,3584,0.009024889104896123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,65536,0.01850488947497474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,1024,0.048807998498280845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,65536,0.06023910972807142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,32,0.009695111049546136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,4096,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,8192,0.3206453323364258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,16384,0.2943742275238037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,1024,0.018400889303949144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,16384,0.1304159959157308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,3584,0.055424888928731285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,65536,0.20930043856302896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,2048,0.013257777525318993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,12288,0.007882666256692674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,10240,0.39150577121310765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,16384,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,12288,0.16062488820817736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,64,0.006680888848172293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,5120,2.0808372497558594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,10240,0.5060177909003364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,5120,0.05801333321465386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,3584,0.20883644951714408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,7168,0.02851466668976678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,5120,0.09914133283827041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,512,0.027657777070999146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,12288,0.014840000205569796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,128,0.011511999699804517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,2048,0.01236533373594284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,1024,0.1167093382941352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,5120,0.04522666666242811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,256,0.0075662218862109715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,51200,1.3912222120496962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,768,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,12288,8.528532240125868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,128,0.005450666778617435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,64,0.006294222341643439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,512,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,51200,0.13935555352105036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,12288,0.7042462031046549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,10240,0.020656888683636982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,5120,0.022966222630606756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,2048,0.024978665841950312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,4096,0.01218488895230823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,128,0.009913778139485253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,3072,0.016383111476898193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,512,0.022082666556040447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,51200,2.6773022545708547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,51200,172.7048882378472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,8192,0.5457839965820312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,65536,41.92698838975694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,5120,0.04257244533962674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,128,0.009490666290124258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,2048,0.02557955516709222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,16384,0.012887111140622033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,3072,0.00888000014755461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,1024,0.4475048912896051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,12288,0.31754665904574925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,2560,0.11310666137271458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,5120,0.12898843818240696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,3584,0.008932444784376357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,512,0.005904888941182031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,12288,0.009007111191749573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,4096,0.023621333969963923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,6144,0.020972443951500788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,65536,0.3072933355967204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,65536,23.674763997395832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,3584,0.012284444438086616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,51200,0.10004888640509711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,4096,0.009921777579519484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,3072,0.03185866607560052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,12288,0.026463111241658527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,3072,0.01055111073785358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,512,0.005609777652555042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,5120,0.3107279936472575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,512,0.0034888887570963967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,1536,0.020429333051045735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,256,0.01477599971824222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,12288,0.7184933556450738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,512,0.0413351125187344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,16384,0.27641956011454266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,8192,0.042262223031785756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,1536,0.009180444810125563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,2560,0.008925333619117737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,512,0.012852444416946836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,1536,0.005707555760939916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,2560,0.026141333911154006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,5120,0.007359999749395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,3072,0.006659555352396435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,10240,0.16104266378614637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,3584,0.007887110941939885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,6144,0.10402933094236587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,65536,0.3810284402635362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,6144,0.11424888504876031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,1536,0.01793066660563151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,10240,0.046738667620552905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,65536,2.351013395521376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,64,0.003012444410059187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,1024,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,7168,0.07646310991711087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,12288,0.39087822702195907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,32,0.002785777880085839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,256,0.003624888757864634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,6144,0.01316177762216992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,16384,0.13430221875508627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,768,0.007473777565691207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,1024,0.004271999829345279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,6144,0.016699555847379897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,1536,0.5878959761725532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,32,0.02652888827853733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,5120,0.48865069283379453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,1536,0.007685333490371704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,768,0.3328719933827718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,12288,0.06308622492684259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,4096,0.00999466660949919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,6144,0.05129422081841362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,1536,0.04984622200330099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,1536,0.00906488878859414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,3584,0.0432897773053911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,3072,0.05102399984995524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,32,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,64,0.006773333168692059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,16384,0.09037244319915771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,2560,0.037250667810440063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,8192,0.02124533388349745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,7168,0.015394666128688388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,768,0.00685422205262714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,256,0.0041546668443414895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,6144,0.03576888971858554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,16384,0.022333333889643352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,65536,1.0371501710679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,3584,0.05586489041646322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,10240,0.015465777781274585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,3584,0.3682897885640462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,8192,0.037053333388434515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,12288,0.02734666731622484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,10240,0.025011556016074285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,768,0.015475556254386902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,32,0.0063733334342638654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,6144,0.1421040031645033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,10240,0.19951644208696154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,16384,2.1595751444498696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,7168,0.04339999953905741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,1024,0.007895111209816402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,2560,0.017824888229370117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,6144,0.08123199807273017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,1024,0.03529244330194261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,4096,0.012910222013791403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,2048,0.007481777833567725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,32,0.002871111035346985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,64,0.002761777697337998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,64,0.0053297777970631914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,65536,0.5779609150356716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,512,0.005543111099137201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,512,0.003703111161788305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,1536,0.5225751135084364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,8192,1.9683297475179036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,51200,0.7565857569376627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,1024,0.013921777407328287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,6144,0.012484444512261285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,12288,0.36974843343098956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,1536,0.01072088877360026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,16384,20.109544542100696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,32,0.0027715555495686005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,10240,0.3291573259565565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,51200,1.1652595731947157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,8192,0.022816888160175745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,768,0.004695999953481886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,512,0.010227555202113258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,3584,0.06919644276301067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,16384,0.06949333349863689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,51200,0.17425955666436088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,2560,0.06989777750439115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,5120,0.017180444465743173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,16384,3.919603559705946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,51200,2.0996969011094833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,3584,0.03273155623012119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,2048,0.1026195552614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,768,0.006593777901596493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,768,0.03711555401484171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,1024,1.365559154086643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,32,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,65536,0.10830310980478923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,51200,0.12809866004519993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,64,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,1024,0.14977422025468615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,7168,0.047114666965272695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,8192,0.08303378025690715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,16384,0.023172444767422144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,3072,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,6144,0.06636089086532593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,16384,0.011904888682895236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,3584,0.01774133410718706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,256,0.050704889827304415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,2560,0.009124444590674506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,6144,0.08759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,3584,0.012527999778588613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,16384,0.4812897576226129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,1024,0.007289778027269576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,51200,4.299432118733724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,2048,0.26591022809346515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,2560,0.014967999524540372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,8192,0.012835555606418185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,2560,0.012128888732857175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,3584,0.011517333487669626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,3072,0.01035733355416192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,1024,0.007916444705592262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,8192,1.6961680518256292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,3072,0.054024888409508594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,2048,0.007085333267847697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,1536,0.010993777877754636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,1024,0.007742221984598372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,512,0.007458666960398356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,65536,0.16254044903649226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,512,0.005566222386227713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,51200,0.8776257832845052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,128,0.06257155868742201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,10240,0.014781332678265043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,16384,1.068336910671658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,10240,0.37242311901516384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,6144,0.006026666611433029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,128,0.004624888714816836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,128,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,3584,0.02383377816942003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,512,0.0038720000949170855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,7168,0.5856595569186741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,7168,0.20623377958933511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,8192,0.038803554243511625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,6144,0.03831111060248481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,128,0.0032026666320032547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,16384,0.37025067541334367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,512,0.0038337777886125776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,4096,0.014730667074521383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,1536,0.016385777129067313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,16384,4.798283471001519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,12288,1.1495253245035808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,65536,0.3767084545559353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,6144,0.028982222080230713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,51200,0.36507201194763184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,16384,1.6054133309258356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,5120,0.028386665715111628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,10240,0.044080889887279935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,8192,0.08190933201048109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,10240,1.2671119901869032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,51200,37.28662448459202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,32,0.00851644410027398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,1536,0.007214222517278459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,5120,0.046839998828040234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,2048,0.09458311398824056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,32,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,16384,0.03333066569434272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,256,0.0032631111227803757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,65536,0.16776711410946318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,4096,0.03236088818973965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,8192,0.02218755582968394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,3072,0.16018311182657877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,3072,0.03720533185535007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,64,0.00703022215101454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,2048,0.007879111501905654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,128,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,2560,0.007730666961934831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,5120,0.03830666674507989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,256,0.0075333333677715724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,2560,0.013249778085284762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,1024,0.005531555662552516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,1024,0.016893333858913846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,6144,0.15086400508880615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,5120,0.25562045309278697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,32,0.0030053332448005676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,1536,0.05132711264822218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,256,0.0029057777590221832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,4096,0.027303111222055223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,12288,0.07931822538375854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,12288,0.02903999884923299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,6144,0.03894755575391982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,1024,0.025936888323889837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,768,0.010213333699438307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,12288,1.2076995637681749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,64,0.14975644482506648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,16384,0.07373422384262085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,3584,0.009400000174840292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,2048,0.005715555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,2560,0.746398237016466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,256,0.013984888792037964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,64,0.03201244605912103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,2048,0.009015111459626092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,2560,0.020041777027977836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,4096,0.012084444363911947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,7168,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,2048,0.01792977750301361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,768,0.0609804458088345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,32,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,5120,0.05486399928728739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,5120,0.1989333364698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,8192,0.012450666891203986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,8192,0.028344889481862385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,5120,0.09435022539562649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,256,0.005517333332035277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,2048,0.007260444263617198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,12288,0.039332442813449435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,5120,0.009384888741705153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,4096,0.1376346614625719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,1024,0.008970666262838576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,128,0.002748444469438659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,2048,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,65536,0.04004977809058295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,512,0.009088888764381409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,2560,0.0726648900243971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,51200,1.3366080390082464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,16384,0.3454719914330377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,2048,0.07578311363855998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,65536,0.3680097791883681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,3072,0.006072000082996156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,1024,0.022894221875402663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,16384,0.04240711198912727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,5120,0.035976890060636736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,7168,0.4680302408006456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,16384,0.03563377923435635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,32,0.009704888694816166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,16384,0.02389777700106303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,3072,0.009963555468453301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,2048,0.00850577818022834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,768,0.04831111099984911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,8192,0.01385600037044949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,32,0.005454222361246745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,10240,1.433892461988661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,256,0.004551110996140374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,64,0.003144888828198115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,3584,0.017104888955752056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,65536,0.15954666667514378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,6144,0.16531199879116482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,768,0.009031111167536842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,5120,0.009202666580677032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,768,0.06533155838648479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,2048,0.005731555736727185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,256,0.030415111117892798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,512,0.005200888961553574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,6144,0.0147706667582194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,512,0.040305776728524104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,256,0.002955555501911375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,3072,0.035396443472968206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,2560,0.009221333596441481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,128,0.005056000004212062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,8192,0.3445484373304579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,10240,0.025518221987618342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,2560,0.06845866971545748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,16384,1.151253382364909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,12288,0.015449777245521545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,4096,0.01921333372592926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,3072,0.013703111145231458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,8192,0.014517333772447375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,10240,0.047381334834628634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,2560,0.017338666650984023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,7168,1.0257270601060655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,256,0.005991999887757831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,65536,0.1000479989581638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,65536,2.4552408854166665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,12288,0.06703733073340522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,6144,0.3146382172902425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,3584,0.1272088951534695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,3584,0.0351164440313975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,7168,0.10512622197469075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,128,0.011757333245542316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,5120,0.09980177879333496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,1024,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,64,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,1024,0.02661866611904568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,512,0.0034604445099830627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,1024,0.029236445824305218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,2048,0.11882755491468643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,2560,0.01495911180973053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,12288,0.08095200194252862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,8192,0.26391998926798504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,2560,0.08379111025068496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,256,0.030504888958401147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,3072,0.010503110786279043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,51200,2.415967093573676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,3072,0.025096888343493145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,8192,0.06377600298987494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,7168,0.0892053312725491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,128,0.00370755543311437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,65536,0.2810257805718316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,512,0.005631999837027655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,768,0.0038933331767717996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,32,0.0026319999661710528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,2048,0.011567999919255575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,256,0.006759111252095964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,768,0.007000889215204451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,256,0.11835643980238174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,7168,0.815608024597168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,7168,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,128,0.004774222357405556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,51200,0.01457244488928053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,6144,0.007319110963079665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,5120,0.0794302225112915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,7168,0.013782222237851886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,4096,0.13441422250535753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,4096,0.01826133330663045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,3584,0.18490933047400582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,512,0.02032177812523312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,768,0.025711110896534387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,256,0.0042568888101312844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,51200,0.1347902218500773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,65536,1.1435937881469727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,3584,0.013811555173661975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,4096,0.012964444855848948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,12288,0.39793509907192653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,16384,0.7148257891337076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,4096,0.013777777552604675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,2048,0.010279110736317104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,12288,0.053336888551712036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,51200,0.18454310629102919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,12288,0.06061600314246284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,10240,0.2102559937371148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,3584,0.5489350954691569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,512,0.010523555179437002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,2560,0.007124444676770105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,256,0.004181333300140169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,1536,0.02281066609753503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,128,0.008007110820876228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,3584,0.1795253356297811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,8192,0.5158978038363987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,6144,0.03193155593342251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,16384,0.02439466615517934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,1536,0.012158222496509552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,128,0.035205334424972534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,1536,0.022443556123309668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,8192,0.030436442957984075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,12288,0.41028179062737363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,128,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,16384,0.10882755782869126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,3072,0.01498044364982181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,64,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,512,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,5120,0.010135999984211391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,16384,0.02106399999724494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,1024,0.006038222048017714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,7168,0.013331555657916598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,7168,0.059897780418395996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,51200,0.07842844724655151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,1024,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,3584,0.09336355659696792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,1536,0.06898044215308295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,2560,0.008488888541857401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,1024,0.006436444405052397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,51200,0.0880497760242886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,8192,0.04718577861785889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,128,0.004412444515360726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,12288,0.2161795563167996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,7168,0.10490666495429145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,51200,0.20185689131418863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,10240,0.03294577863481309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,6144,0.04560977882809109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,4096,0.020301333732075162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,768,0.010696888797812991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,4096,0.21264266967773438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,64,0.008577778107590145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,8192,0.03327555457750956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,1024,0.03528533379236857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,51200,1.7665706210666234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,1536,0.1363582213719686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,2560,0.03028622269630432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,1536,0.0046693334976832075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,51200,0.02932000160217285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,6144,0.015650666422314115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,7168,0.04144710964626736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,51200,0.15385244952307806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,3072,0.005988444305128521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,3072,0.05305599835183886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,4096,0.40667377577887637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,768,0.08498577939139472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,3072,0.009362666971153682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,10240,0.03420977791150411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,2048,0.08694310983022054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,3584,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,32,0.005136000199450387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,51200,5.28980975680881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,51200,2.6249190436469183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,7168,0.016693333784739178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,8192,0.0819831093152364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,10240,0.8247982131110297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,1536,0.09352355533175999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,5120,0.1542524496714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,12288,0.07983111010657416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,7168,0.1642871167924669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,65536,0.9031706915961372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,1024,0.29553598827785915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,128,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,3072,0.01699822147687276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,2048,0.015856888559129503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,6144,0.03755911191304525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,128,0.009667555491129557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,64,0.005662222289376789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,256,0.0058808889653947614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,64,0.0028826666788922418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,64,0.0052382221652401825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,1024,0.2665777736239963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,12288,0.02583733366595374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,32,0.01716977854569753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,51200,0.1921182208591037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,4096,0.010618666807810465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,2048,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,8192,2.976366255018446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,51200,0.21223555670844185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,3584,0.05164444446563721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,32,0.007220444579919179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,51200,1.3643804126315648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,1536,0.006821333534187741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,7168,0.014352889524565803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,512,0.005543111099137201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,4096,0.010649777948856354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,64,0.0029111111329661477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,128,0.0034266666819651923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,65536,3.2150452931722007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,32,0.01036177741156684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,256,0.02125955621401469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,51200,1.3485191133287218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,1024,0.03270755542649163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,512,0.008279999925030602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,2048,0.03280177712440491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,7168,0.035411556561787925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,3072,0.0638435549206204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,5120,0.007304888632562425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,2048,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,3072,1.8377742767333984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,2560,0.01806400054030948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,2048,0.012454222473833295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,5120,0.026759111218982275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,64,0.009289777941173976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,32,0.022686221533351477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,2560,0.031926221317715116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,6144,0.45331822501288516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,8192,0.04026222229003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,12288,6.0014088948567705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,4096,0.0727635555797153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,2048,0.0570346646838718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,256,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,2560,0.044176889790429004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,128,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,768,0.015480000111791821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,1024,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,1024,0.013032000097963544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,768,0.008767111433876885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,65536,0.014479999740918478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,2560,0.0063031112982167145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,65536,0.4128257698482937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,64,0.005474666754404704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,768,0.008644444247086843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,2560,0.00654666663871871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,2048,0.006400888992680444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,12288,0.015185778339703878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,128,0.01872177753183577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,3584,0.010658666491508484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,3072,0.005921777751710679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,51200,2.129331588745117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,65536,0.08760622474882339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,3584,0.2279555532667372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,512,0.015264888604482016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,512,0.025349333882331848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,32,0.007740444607204861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,3072,2.6468817392985025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,256,0.008568888737095727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,6144,0.012873777912722694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,2048,0.027429333991474573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,128,0.03095199995570713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,2048,0.011612444288200803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,16384,0.17216000292036268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,3072,0.008244444098737504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,7168,0.015277332729763456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,2560,0.00869777798652649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,10240,0.017847999930381775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,256,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,768,0.0037226668662495087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,16384,1.8494391971164281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,51200,0.3537404537200928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,3072,0.007965332931942409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,1024,0.010015111002657149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,65536,0.03228888909022013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,128,0.005998222364319696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,3584,0.008982222113344405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,1536,0.06159910890791151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,6144,0.021692444880803425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,7168,0.05382844474580553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,64,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,5120,0.011634666886594562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,768,0.19895733727349174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,16384,0.060481780105166964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,128,0.012688889271683164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,16384,0.036335110664367676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,12288,0.02770044406255086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,6144,0.1355537838406033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,5120,0.015648888217078317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,65536,0.21942667166392008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,3072,1.3010578155517578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,12288,0.09504533476299709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,64,0.003516444315512975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,1024,0.005070222334729301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,12288,0.01626133256488376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,512,0.00792622235086229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,5120,0.040222220950656466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,4096,0.30991821818881565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,1024,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,768,0.011023999916182624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,2560,0.022666666242811415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,2048,0.023810666468408372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,2048,0.006058666855096817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,6144,0.015986666083335876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,768,0.03542133172353109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,10240,0.07066666417651706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,7168,0.10713333553738064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,8192,0.04763466782040066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,51200,0.0875848862859938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,10240,0.8659768634372287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,768,0.35437954796685117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,64,0.010984888507260216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,64,0.012973333398501078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,7168,0.22850312127007377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,3072,0.030621333254708186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,512,0.030864887767367895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,5120,0.024209777514139812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,51200,0.18132266733381483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,1536,0.006599999964237213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,12288,0.01391377713945177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,16384,0.4765431086222331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,6144,0.11286844147576226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,512,0.005039999882380168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,32,0.013615111509958902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,16384,0.14662133322821722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,4096,0.012260444462299347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,10240,0.01949155496226417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,65536,0.16099554962582058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,1536,0.010046222143703038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,16384,0.2902195453643799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,2560,0.06969600253634982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,3072,0.01220444424284829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,2560,0.24040799670749238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,65536,0.10734222332636516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,8192,0.01928622192806668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,3072,0.4250071048736572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,256,0.0058062221441004015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,16384,0.2987857659657796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,8192,0.08268621895048353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,8192,0.37791911760965985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,32,0.002842666581273079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,51200,0.8829253514607748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,3072,0.009149333669079674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,65536,0.4108097818162706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,5120,0.032792889409595065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,16384,0.5502790874905056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,6144,0.030831111801995173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,3072,0.009391110804345872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,16384,0.29042667812771267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,10240,0.02560088866286808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,768,0.008760888543393876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,8192,0.016921778519948322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,6144,0.010220444036854638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,256,0.009723555710580613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,51200,0.19382845030890572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,1024,0.01973777843846215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,65536,0.1771226723988851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,6144,0.019497777024904888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,2048,0.00811111099190182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,1024,0.011928888658682505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,4096,0.011870221959220039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,2560,0.3472844494713678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,5120,0.09315022495057847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,32,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,128,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,128,0.006516444600290722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,16384,0.36693334579467773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,51200,0.23050665855407715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,128,0.005295111073387994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,3584,0.09177866909239028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,512,0.005447111195988125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,2048,0.04718755682309469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,51200,0.007967111137178209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,4096,0.19577511151631674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,1536,0.00463733325401942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,5120,0.03763733307520548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,16384,0.27724266052246094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,1024,0.8007689052157931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,5120,1.4792888429429796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,2048,0.016528889536857605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,8192,0.02852533260981242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,6144,0.005945777727497949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,3584,0.03009066647953457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,51200,0.8628160158793131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,768,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,1536,0.03162844313515557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,4096,0.0169493340783649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,8192,0.036115556955337524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,5120,0.011295999917719098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,6144,0.005872888697518243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,4096,1.5509955088297527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,2560,0.021945777866575453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,3584,0.009223110973834991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,3072,0.011534222298198275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,65536,0.3320479922824436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,7168,0.02327377763059404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,1024,0.011304000185595619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,12288,0.3385946750640869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,512,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,6144,0.45984352959526914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,2560,0.007357333269384172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,512,0.03496177660094367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,3584,0.036997334824668035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,6144,0.11973867151472305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,128,0.05475111140145195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,16384,0.022637334134843614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,5120,0.12096977233886719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,7168,0.15682400597466364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,10240,0.4082640012105306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,2560,0.12016267246670193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,768,0.0058346668051348785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,1536,0.01676266723208957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,8192,0.2638755639394124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,4096,0.02843733297453986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,6144,0.2172000010808309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,65536,0.10550310876634385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,6144,0.012437333663304647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,256,0.005076444397370021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,128,0.004289777742491828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,7168,0.0797262191772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,1536,0.0063582224150498705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,768,0.30136532253689235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,4096,0.09308444129096137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,10240,0.06534488995869954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,10240,0.42998311254713273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,8192,1.4786355760362413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,8192,0.027615111735132005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,12288,0.37400444348653156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,65536,0.04788000053829617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,1536,0.02268444498380025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,1536,0.009557333257463243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,51200,1.8349225785997179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,1536,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,3584,0.05854933129416572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,6144,0.01995733380317688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,16384,0.7458657688564725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,7168,0.021743999587164983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,2048,0.01725955473052131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,7168,0.11368710464901394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,2048,0.006887110984987683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,51200,0.06281955374611749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,4096,0.030757334497239854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,32,0.004265777766704559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,10240,0.0920613341861301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,768,0.005838222387764189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,8192,0.26219020949469674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,512,0.008632000121805403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,3584,0.011598221957683563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,10240,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,2048,0.1293297741148207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,3584,0.009118222528033787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,3072,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,256,0.00637511122557852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,128,0.002797333316670524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,12288,0.12115911642710368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,7168,0.02873777680926853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,51200,1.2896284527248807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,16384,2.3280603620741105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,6144,0.01885600056913164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,32,0.006245333287451003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,2560,0.07638311386108398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,32,0.0032106666929192017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,256,0.03514044483502706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,512,0.00685511115524504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,1024,0.029192888074451025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,512,0.06739199823803373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,7168,0.038924445708592735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,7168,0.0867875549528334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,2048,1.5745902591281469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,8192,0.022848000129063923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,65536,0.09313155545128717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,3584,0.2071875598695543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,4096,0.018557333283954196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,7168,0.15408178170522055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,3072,0.0075866662793689305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,1536,0.011505777637163797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,12288,0.024846222665574815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,3072,0.02344266739156511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,7168,0.020941333638297185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,51200,0.0894213318824768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,4096,0.06101600329081217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,4096,0.010024888647927178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,16384,0.28176267941792804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,768,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,1024,0.006169777777459886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,128,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,8192,0.4434071116977268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,4096,0.012512000070677863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,512,0.03530044356981913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,768,0.032666666640175715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,6144,0.01626755628320906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,128,0.008195555872387355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,64,0.005168000029193031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,2048,0.04335822330580818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,65536,0.6525680224100748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,512,0.006462222172154321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,3072,0.07148533397250705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,3584,0.010100444157918295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,768,0.29395821359422475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,65536,0.2496088875664605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,51200,0.08948355250888401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,12288,0.021578666236665513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,5120,0.009838222629494136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,7168,0.1407688856124878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,5120,0.010852444503042432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,51200,0.00963644435008367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,2048,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,7168,0.17654221587710908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,2560,0.0173342227935791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,256,0.04684177703327603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,12288,0.018814222680197824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,16384,9.659114413791233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,16384,0.02689244515366024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,1536,0.5115840170118544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,768,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,2560,0.01036888857682546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,512,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,65536,2.4445554945203996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,7168,0.2237671216328939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,1024,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,1536,0.03145422206984626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,2048,0.010025777750545079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,768,0.008752889103359645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,3072,0.07705333497789171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,65536,4.336656782362196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,2048,0.06397689051098294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,256,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,6144,0.010267555713653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,1024,0.011716444459226398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,256,0.004405333350102107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,6144,0.04642400145530701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,512,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,1024,0.09944266743130153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,65536,3.6587927076551647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,64,0.042635556724336415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,65536,0.921536021762424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,512,0.682710223727756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,64,0.010746666954623329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,256,0.003683555457327101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,1536,0.16411111089918348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,4096,0.17813155386182997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,128,0.01354311075475481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,51200,0.3731999927096897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,512,0.00498577786816491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,4096,2.8866329193115234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,1536,0.11130577988094753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,65536,1.0309324264526367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,512,0.005995555470387141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,64,0.0041582224269707995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,4096,0.010864888628323873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,10240,0.06901688708199395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,7168,0.007811555431948767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,5120,0.04974400003751119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,51200,0.2892728911505805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,51200,0.22910934024386934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,1024,0.011715555356608497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,2048,0.0440248880121443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,32,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,51200,0.0505600008699629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,10240,0.0394702222612169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,512,0.00721777809990777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,1536,0.15649955802493626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,8192,0.03909511036343045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,768,0.005147555636035072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,768,0.0047591109242704184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,65536,0.3217253420088026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,256,0.0037751110891501107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,12288,3.4037715064154734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,32,0.010512000156773461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,3072,0.1463093360265096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,10240,0.01979111135005951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,128,0.021147555775112573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,8192,0.008387555678685507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,7168,0.02456088860829671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,128,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,10240,0.0446631113688151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,256,0.009520000053776635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,2560,0.020975111259354487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,2560,0.01092355532778634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,51200,1.7559164894951715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,64,0.007469333708286285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,7168,1.4225484000311956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,7168,0.10957155625025432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,16384,0.22948087586296928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,16384,0.027181333965725366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,768,0.003622222277853224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,128,0.01828444500764211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,51200,1.1610950893825955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,4096,0.06810844606823392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,10240,0.19004178047180176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,10240,0.02641422218746609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,10240,1.4913884268866644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,16384,0.03649422195222642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,3072,0.18396355046166313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,1536,0.012588444683286877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,1024,0.015240000353919135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,10240,0.2268995443979899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,51200,0.08490133285522461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,7168,0.03026755650838216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,10240,0.13941421773698595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,1536,0.01163999984661738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,4096,0.032468444771236844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,10240,0.06406399938795301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,128,0.0037457777394188773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,2560,0.016629333297411602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,16384,0.006200888918505774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,10240,0.017210667332013447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,1024,0.007058666812049017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,5120,0.04121333360671997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,256,0.01366133325629764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,8192,0.030344890223609075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,5120,0.008841777841250101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,16384,0.18047022819519043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,2048,0.005410666680998272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,512,0.005561777700980504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,65536,33.04604763454861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,256,0.03591377867592706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,10240,0.14628177218967015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,7168,0.02312533391846551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,16384,0.029040889607535467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,512,0.0351875556839837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,6144,0.11008799738354152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,64,0.01687022215790219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,7168,33.19379001193577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,8192,0.11986755000220405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,6144,0.2849280039469401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,12288,0.049032890134387545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,12288,0.23904000388251412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,3584,0.03690133492151896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,128,0.004085333396991094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,6144,0.08278044727113512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,1536,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,256,0.005333333379692501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,128,0.08798044257693821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,2560,0.00977244476477305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,2048,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,3072,0.1594124502605862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,4096,0.02605155607064565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,65536,0.10938311285442776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,2560,0.03086400032043457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,128,0.05107466710938347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,1024,0.0047191112405723995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,5120,0.008803555534945594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,2560,0.04989333285225762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,2560,0.008398222426573435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,3072,0.010106667048401302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,8192,0.01834133267402649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,1536,0.14385778374142116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,256,0.00740533322095871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,51200,0.372416893641154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,32,0.14835911326938206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,12288,0.08487911356820001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,5120,2.044658660888672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,1536,0.008647999829716152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,3584,0.02784444557295905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,1024,0.007520889242490132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,4096,0.286972443262736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,65536,1.5534355375501845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,4096,0.013369777964221107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,3584,0.010061333576838175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,512,0.003395555540919304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,51200,0.07824444770812988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,51200,0.5217910872565376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,1024,0.007456888755162557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,256,0.00620888877246115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,6144,0.02310933338271247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,5120,0.06596533457438152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,65536,1.666589313083225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,51200,0.10846400260925293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,1024,0.018406222263971966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,3072,0.006412444429265128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,3072,0.006248888870080312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,16384,0.0218551109234492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,5120,0.008373333348168267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,3584,0.06632177697287665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,768,0.009114666945404476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,16384,0.053242666853798755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,10240,0.01789777808719211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,2560,0.04213866591453552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,10240,0.01236177815331353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,51200,0.06329244375228882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,7168,0.028031110763549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,51200,0.2130133310953776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,16384,0.35341421763102215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,3584,0.011334222224023608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,64,0.003632889025741153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,4096,2.4866809844970703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,4096,0.24234755833943686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,8192,0.02700177828470866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,2560,0.17860621876186797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,2560,2.092502170138889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,51200,0.07162755727767944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,256,0.005444444302055571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,2560,0.01015911085738076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,768,0.041371554136276245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,128,0.004908444566859139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,6144,0.2710142135620117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,512,0.00582755563987626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,512,0.004993777722120285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,2048,0.018744889232847426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,768,0.0039057777159743835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,6144,0.012520888613329994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,768,0.023577777875794306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,4096,1.0414951112535265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,10240,0.037680887513690524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,256,0.0046471113132105935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,10240,0.04514133267932468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,10240,0.1063413355085585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,4096,0.02381955583890279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,3584,0.06202933523390028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,1024,0.259789334403144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,2560,0.040688001447253756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,3072,0.01770755483044518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,3072,0.23339645067850748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,3584,0.3627280129326715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,256,0.007772444850868649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,6144,0.23072266578674316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,16384,0.06976799832450019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,2560,0.057368887795342334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,7168,0.023847111397319373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,2560,0.021162667208247717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,768,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,2560,0.020069332586394418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,768,0.10632622241973877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,51200,4.244635687934028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,65536,0.22835999064975313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,512,0.008660444782839881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,5120,0.02068000038464864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,8192,0.058964444531334766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,768,0.00462666650613149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,6144,0.06437155273225573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,4096,0.12647910912831625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,3584,0.03338755501641168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,10240,0.018396443790859647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,10240,18.159688313802082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,10240,0.035432888401879206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,3584,0.19441866874694824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,512,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,51200,0.5710568957858616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,2560,0.05800888935724894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,65536,3.1829831865098743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,2048,0.024345777100986902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,12288,0.007868444754017724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,3072,0.016887111796273124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,5120,0.09090399742126465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,1024,0.008588444855478074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,2048,0.0520648890071445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,256,0.008613333106040955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,3584,0.012039999994966718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,64,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,2048,0.009679111341635386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,65536,0.2436453236473931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,2048,0.006141333530346553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,1536,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,12288,0.027170666389995154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,4096,0.3016924328274197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,4096,0.2964364422692193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,51200,0.27727821138170033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,1536,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,16384,3.3913928137885203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,6144,0.010372444159454769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,512,0.004517333375083076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,256,0.024013333850436743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,51200,0.128602663675944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,1536,0.014161777165200977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,16384,0.03167644474241469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,2048,0.020617778102556866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,2048,0.0074444446298811175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,3072,0.00904177791542477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,12288,0.7751190927293566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,65536,0.07295111152860853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,128,0.0034560000316964257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,5120,0.019130667050679524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,2048,0.007424888511498769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,12288,0.006711999989218182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,65536,0.23658667670355904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,65536,0.31880177391899955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,7168,0.21236356099446616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,6144,0.019077334139082167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,2048,0.008192888564533657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,4096,0.019900444481107924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,512,0.030065778228971694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,768,0.01809866726398468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,5120,0.5631422466701931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,5120,0.07521866427527533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,10240,2.5325582292344833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,5120,0.013756444056828817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,2048,0.006842666616042455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,51200,1.3086960050794814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,1024,0.004232888834344016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,6144,0.021738666627142165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,2048,1.100006209479438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,2560,0.880000909169515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,16384,0.180275559425354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,256,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,1024,0.8797013494703504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,1536,0.029285334878497656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,32,0.012473777764373355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,4096,0.01517955627706316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,128,0.0063573333124319715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,10240,0.053872889942593045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,16384,0.02128711177243127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,3584,0.02211644417709774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,51200,0.22163821591271293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,64,0.014694222145610385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,1024,0.01055022246307797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,4096,0.014434667097197639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,12288,0.33295022116767037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,64,0.004596444467703502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,12288,0.02546577817863888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,4096,0.06515822145673963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,64,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,3072,0.16862755351596406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,3584,0.11006222168604533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,2560,0.01147911118136512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,3584,0.024719999896155462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,3072,0.14409777853224012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,2560,0.08465777503119574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,10240,0.17587199476030138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,256,0.026349332597520616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,768,0.012183999849690331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,10240,0.006380444599522485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,65536,0.08642399973339504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,2048,0.017479111750920612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,768,0.003702222059170405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,6144,0.15872711605495876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,3584,0.010688888529936472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,32,0.10613955391777886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,5120,0.7022968928019205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,2048,0.09211377965079413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,8192,0.037295109695858426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,2560,0.04408355553944906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,256,0.0069546666410234236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,32,0.010523555179437002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,768,0.036044445302751325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,65536,0.19347643852233887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,128,0.003967111309369405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,16384,6.174609714084202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,2048,0.010239111052619087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,512,0.0033928888539473214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,5120,0.029189334975348577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,7168,1.5199973848130968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,7168,0.03187288840611776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,128,0.0027884443600972495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,7168,0.01909955508179135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,10240,0.2148275507820977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,10240,0.7656284438239204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,512,0.009337777892748514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,3072,0.016422222057978313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,10240,0.04265244470702278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,2048,0.01720888912677765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,5120,0.033450666401121355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,8192,0.01536177761024899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,2560,0.01678311162524753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,1536,0.011952000359694162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,512,0.005842666659090254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,51200,0.1255777809354994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,32,0.0026453334010309647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,12288,0.09022311369578044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,16384,0.13076178232828775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,512,0.006690666493442323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,128,0.15882489416334364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,51200,0.13953777154286703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,1024,0.006628444625271692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,128,0.16018933720058864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,16384,0.011866666376590729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,6144,0.02032355633046892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,512,0.11038044426176284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,5120,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,1024,0.007017778025733099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,3584,0.0122079998254776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,768,0.025808889004919264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,10240,0.042768001556396484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,5120,0.006680000159475539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,2048,0.007183111376232571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,65536,3.3502943250868054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,65536,4.116341484917535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,2560,0.1870488855573866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,51200,2.5822230445014105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,51200,0.1849679946899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,2048,0.01754755609565311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,65536,0.5069066683451334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,65536,0.13134045071072048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,512,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,1024,0.05075022247102526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,256,0.005055111315515307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,5120,0.014690667390823364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,16384,0.022156443860795762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,65536,0.9122346242268881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,2560,0.32085956467522514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,10240,0.02546666728125678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,512,0.019984000258975558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,65536,0.2493520047929552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,10240,0.014557333456145393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,7168,0.029205332199732464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,8192,0.027787556250890095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,3584,0.008159110943476358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,10240,5.043517218695746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,5120,0.01367288910680347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,2048,0.014698666003015308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,12288,0.04349244303173489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,512,0.010584889186753167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,16384,0.059875554508633085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,1536,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,3584,0.018262222409248352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,3072,0.020727111233605277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,16384,0.0144533332851198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,512,0.005934222290913264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,3072,0.5434702237447103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,512,0.10848000314500596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,128,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,51200,0.11820444795820449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,4096,0.46667109595404727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,256,0.027626666757795546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,51200,0.08443199925952488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,2560,0.010282666318946415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,1536,0.00656444455186526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,6144,0.06404444244172838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,3072,0.05566222137875027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,3072,0.15575022167629665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,65536,0.047041777107450694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,768,0.005853333406978183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,4096,0.009604444106419882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,3072,0.8491466840108236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,4096,0.08838844299316406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,64,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,32,0.008291555775536431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,1024,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,3584,0.017535999417304993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,2560,0.011634666886594562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,65536,0.27882311079237193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,256,0.7211893399556478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,768,0.04178577661514282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,4096,0.0303111109468672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,2560,0.01903377804491255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,2560,0.03952266772588094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,3072,0.008779555559158325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,2048,0.013215999636385175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,3584,1.1106221940782335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,5120,0.028174221515655518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,2560,0.03636888994110955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,3072,0.23462398846944174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,2048,0.0432240002685123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,6144,0.018144889010323417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,128,0.00463733325401942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,6144,0.40734044710795086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,10240,0.0262435558769438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,3584,0.0069324444565508105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,512,0.007263110743628607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,16384,0.1979315545823839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,128,0.01346577786737018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,2048,0.028459555572933618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,5120,0.02126755482620663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,1536,0.00675644435816341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,4096,0.01125600023402108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,1536,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,4096,0.11360889010959202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,65536,0.6613981988694932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,768,0.009901333186361525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,4096,0.017312000195185345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,16384,0.07550488577948676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,768,0.007264888948864407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,5120,0.11211466789245605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,2560,0.021182222498787776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,16384,0.024903110331959192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,3072,0.011869333684444427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,64,0.00489777781897121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,51200,0.37403021918402773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,512,0.04067644476890564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,10240,0.06804888778262667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,12288,0.019934222102165222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,1024,0.010444444086816577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,2560,0.6368853251139323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,6144,0.010795555180973478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,768,0.020536889632542927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,256,0.005606222069925732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,2048,0.11694310771094428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,6144,0.018491556247075398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,128,0.003113777687152227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,4096,0.012053333222866058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,64,0.035951998498704694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,3584,0.0710737771458096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,768,0.023856888214747112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,3072,0.01774311065673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,5120,0.7490257687038846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,2048,0.007742221984598372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,12288,0.12670044104258218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,2560,0.09956444634331597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,512,0.017330666383107502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,65536,0.26814132266574436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,512,0.023004444109068975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,7168,0.007927999728255803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,1024,0.013839999834696451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,4096,0.02088711162408193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,128,0.004504888835880491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,128,0.004544888933499654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,3584,0.26768355899386936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,12288,0.1879662275314331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,768,0.006236444330877728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,32,0.0064284445510970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,7168,0.026328888204362657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,4096,0.025632888078689575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,2560,0.18561689058939615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,256,0.013423110875818463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,10240,0.03083911206987169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,12288,0.025052444802390203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,65536,2.463663101196289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,256,0.037660443120532565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,3584,0.03162666824128892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,16384,0.10066399971644084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,10240,0.04341244366433886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,32,0.005314666777849197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,512,0.4104311201307509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,1024,0.0074542222751511475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,128,0.002944888960984018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,12288,0.020488889680968392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,3072,0.08698844247394139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,3584,0.013612444202105204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,7168,0.02215555475817786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,768,0.23952976862589517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,2048,0.10802577601538764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,2048,0.01440000037352244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,768,0.020079111059506733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,5120,0.019469334019554984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,16384,0.02901600135697259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,16384,0.01388533330625958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,16384,0.42926843961079914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,4096,0.03745155533154806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,1536,0.022973333795865376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,256,0.012911111116409302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,8192,0.1877279943890042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,1536,0.01685599982738495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,32,0.010125333236323463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,3072,0.0064071110553211635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,65536,0.212954666879442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,2048,0.010135999984211391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,256,0.007793777518802219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,8192,0.023759111762046814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,2560,0.025212445192866858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,256,0.006817777951558431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,65536,0.12115911642710368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,256,0.005088888936572605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,1024,0.03578044308556451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,64,0.008171555896600088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,512,0.010888888604111142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,1024,0.0074035558435651995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,10240,0.013754666679435305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,768,0.02350133326318529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,16384,0.01868977811601427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,1024,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,1536,0.03271911210483975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,5120,0.021785777476098802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,256,0.15657156043582493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,2560,0.007151111132568783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,512,0.007340444458855524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,12288,0.05750844213697645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,6144,0.023502222365803186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,2560,0.1738711065716214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,3072,0.01163733336660597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,12288,0.041738665766186185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,10240,0.02762577765517765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,7168,2.465651618109809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,3072,0.014462222655614218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,65536,0.12869332896338567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,12288,0.031952887773513794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,10240,0.13926666312747532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,10240,0.22460444768269858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,6144,0.02833066715134515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,16384,0.26988087760077584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,3072,0.06588000059127808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,65536,0.06804711288876004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,7168,0.47135199440850156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,16384,0.14837688869900173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,5120,0.1135937770207723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,512,0.005664000080691443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,512,0.01537333263291253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,7168,0.05096444487571716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,8192,0.01185955521133211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,6144,0.04530310961935255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,5120,0.0207368897067176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,12288,4.277112748887804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,65536,0.04697511262363858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,8192,0.01668355531162686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,32,0.0617048872841729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,8192,0.09166755278905232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,32,0.010268444816271463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,10240,0.05625155236985949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,51200,0.23519023259480795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,5120,0.09000711308585273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,512,0.006048000107208888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,128,0.005528888768619961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,256,0.008485332959228093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,256,0.04955910974078708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,65536,0.12653599845038518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,256,7168,0.032396445671717324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,65536,39.04869588216146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,16384,0.03232533401913113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,7168,0.170270217789544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,64,0.003101333354910215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,1536,0.01629866659641266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,51200,0.9401093588935004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,32,0.024223110742039148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,51200,4.2411723666720915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,3072,0.023443554838498432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,512,0.004561777744028303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,4096,0.04937244455019633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,16384,0.16281244489881727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,768,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,5120,0.021551110678248938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,2048,0.0321395562754737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,4096,0.07078933053546481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,3584,0.04281155599488152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,512,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,3584,0.00796622203456031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,512,0.004906666775544484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,10240,0.017495110630989075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,4096,0.2892017894321018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,1024,0.0069955554273393415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,2560,0.0068959999415609576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,16384,0.03548355566130744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,32,0.03297777639495002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,5120,0.028918223248587713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,7168,0.3513333267635769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,8192,0.04175644450717502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,12288,0.4031199879116482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,2048,0.005883555445406172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,64,0.006511111226346757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,3072,0.01275022245115704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,10240,0.026668444275856018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,16384,4.477024926079644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,1024,0.0047386665311124595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,3584,0.036171555519104004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,51200,0.054682665401034884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,5120,0.05858844518661499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,3072,0.00964355551534229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,2560,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,8192,0.016912889149453905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,256,0.003371555565132035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,6144,0.009092444347010719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,7168,0.02126311096880171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,5120,0.022335110439194575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,16384,0.009379555781682333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,256,0.01329066687160068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,512,0.0041146667467223275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,3584,0.03895733422703213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,7168,0.021846221552954778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,2048,0.0064479998416370815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,3584,0.02558311157756382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,1024,0.0123511114054256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,8192,0.04210755560133192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,8192,0.17784889539082846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,3584,0.17063377963172066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,1536,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,10240,0.07194488578372531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,256,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,6144,0.018941332896550495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,32,0.009555555880069733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,512,0.034895999564064875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,7168,0.014335999886194864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,4096,0.020780444145202637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,256,0.009979555176364051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,3584,0.013647110925780402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,2048,0.017086222767829895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,6144,0.054881778028276235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,5120,0.031311111317740545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,65536,0.014109333356221518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,2048,0.02411199940575494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,5120,0.04333510994911194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,3072,0.02422933280467987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,5120,7.832526312934028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,4096,0.014132445057233175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,8192,0.023593778411547344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,1536,0.009002666506502364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,65536,2.0880283779568143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,1024,0.004835555536879433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,10240,0.28952622413635254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,7168,0.09898933437135483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,1024,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,2048,0.013912000589900546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,7168,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,12288,0.021130666136741638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,768,0.025480889611774023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,128,0.015397333436542086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,12288,0.04745066828197903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,3072,0.17109421888987222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,12288,0.021560889151361253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,16384,0.2472595638699002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,65536,0.06724266873465644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,128,0.003801777958869934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,5120,0.017154667112562392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,12288,0.04215733210245768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,6144,0.012704888979593912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,3584,0.0548622210820516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,6144,0.010837333069907295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,12288,0.0433884428607093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,1536,0.005709333138333426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,768,0.008635555704434713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,8192,0.63064177831014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,512,0.005697777701748743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,10240,0.00794044468137953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,3072,0.12998578283521864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,32,0.007664889097213745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,10240,0.03785511189036899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,3584,0.03590577840805054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,2048,0.015265777707099915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,7168,0.14128000206417507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,2560,0.02216533323129018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,4096,0.03587377733654446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,256,0.01292711082432005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,512,0.007649777664078607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,16384,0.16347466574774847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,5120,0.03170399864514669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,4096,0.14297866821289062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,2048,0.08567466338475545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,32,0.002992000016901228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,1024,0.006047111004590988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,5120,0.050139556328455605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,128,0.004436444491147995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,32,0.004655999855862724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,16384,0.6729235649108887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,1536,0.060039109653896756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,65536,33.064551459418404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,12288,0.18871822622087267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,256,0.02318222158484989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,10240,0.10017600324418809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,3584,0.013533333109484779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,32,0.0026826666047175727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,512,0.016224000189039443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,64,0.002708444371819496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,7168,0.02779200010829502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,2048,5.993684556749131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,16384,0.025111999776628282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,12288,0.17431467109256318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,3584,0.06782666842142741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,1024,0.0058293334311909145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,10240,0.037188443872663714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,1536,0.0604577793015374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,5120,0.49492624070909286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,3584,0.08979466888639663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,32,0.006654222392373615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,1024,0.04014399978849623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,6144,0.01628800067636702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,1536,0.022480888499153986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,768,0.021687999367713928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,3072,0.06274488899442884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,1536,0.016872889465755887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,512,0.013101333545313941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,3584,0.04869777626461453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,65536,0.4814880159166124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,6144,0.3235066731770833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,3072,0.041152887874179415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,16384,0.019736889335844252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,7168,0.015093333191341825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,3072,0.028051555156707764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,2560,0.15199111567603216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,3584,0.08480355474683975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,256,0.005145777844720417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,256,0.005247999810510212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,2048,0.028055999014112685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,64,0.006542222367392645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,51200,0.047678222258885704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,51200,0.05307022068235609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,8192,0.21276888582441542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,10240,0.023071999351183575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,1536,0.009687110781669617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,12288,17.089786105685764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,768,0.05513599846098158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,5120,0.011884444289737277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,7168,0.19810222254859078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,5120,0.012210666305489011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,8192,1.233965343899197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,4096,0.037125332487954035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,256,0.0035137778355015647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,128,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,6144,0.01694222291310628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,12288,0.1305511130226983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,6144,0.09251022338867188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,2560,0.3641768826378717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,32,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,4096,0.01744622157679664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,2560,0.0059164443777667145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,5120,0.09117244349585639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,3072,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,5120,0.009317333499590555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,10240,0.022499556342760723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,65536,1.5737333297729492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,128,0.004506666627195147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,128,0.008657777474986183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,256,0.050160888168546885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,7168,0.1603297789891561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,64,0.012070222033394707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,256,0.0058471109304163195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,32,0.011258666714032492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,16384,0.2820817894405789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,64,0.0027022223091787765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,5120,0.009720000127951304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,3072,1.7627653545803492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,51200,13.35821533203125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,7168,0.00792622235086229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,10240,0.13610933886633977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,8192,0.006801777829726537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,16384,0.9464595582750109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,3584,0.025569778349664476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,2560,0.023435556226306494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,256,0.011204444699817233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,128,0.005290666802061929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,12288,0.014442667365074158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,10240,0.013654222091039022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,5120,0.008597333398130205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,6144,0.012368000215954252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,16384,0.0352151095867157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,4096,0.22417600949605307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,8192,0.01701244380739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,65536,0.6625439855787489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,32,0.003112888791494899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,1024,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,16384,0.22461512353685167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,51200,0.07212533553441365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,64,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,3072,0.04873955580923292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,256,0.020185778538386028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,51200,0.18966755602094862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,512,0.25989243719312877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,2560,0.014138667119873894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,2560,0.011226666470368704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,2048,0.006990222467316522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,65536,5.734233856201172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,5120,0.2269066704644097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,1536,0.006254222244024277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,51200,0.07353244225184123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,65536,0.07238933112886217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,1024,0.05836177534527249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,768,0.021193777521451313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,512,0.05234577920701769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,768,0.011171555353535546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,1024,0.03498488995763991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,3584,0.10962666405571832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,3584,0.1394044425752428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,2048,0.014589332871966891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,64,0.003195555466744635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,10240,0.08519288566377427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,5120,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,512,0.06977955500284831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,65536,0.5409484439425999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,5120,0.14950044949849448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,6144,0.009860444400045607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,65536,0.6132782300313314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,128,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,65536,0.45574310090806747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,65536,0.07981955342822604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,12288,0.16822400357988146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,2048,0.005780444376998478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,1024,0.005728888842794631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,5120,0.053210665782292686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,1024,0.0041866666740841335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,768,0.00600622221827507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,768,0.009015999734401703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,3584,0.8129946920606824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,128,0.1137839953104655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,32,0.01130222198035982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,2048,0.1688337723414103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,16384,0.10498222377565171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,2048,0.01660977800687154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,5120,0.018219555417696636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,1024,0.02381244467364417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,16384,0.03477688961558872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,2560,0.06953955358929105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,64,0.003951999876234266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,64,0.0225911107328203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,12288,0.017608000172509086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,1024,0.010270222193664974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,3584,0.020597333709398907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,512,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,5120,0.014018666413095264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,512,0.004893333133724001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,256,0.004804444395833545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,16384,0.04925510949558682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,8192,0.4047866662343343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,1536,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,16384,0.08051911327573988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,51200,3.0262408786349826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,7168,0.26830932829115123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,1536,0.010614222122563256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,8192,0.011503111157152386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,65536,0.3100399971008301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,512,0.03395377927356296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,128,0.008239111138714684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,12288,0.03417155477735732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,10240,0.03834222091568841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,4096,0.013295999831623502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,12288,1.1910346349080403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,1024,0.16279289457533094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,3584,0.011204444699817233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,12288,0.021764443980322942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,8192,0.25413333045111763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,8192,0.060961776309543185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,32,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,1024,0.007768889268239339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,3584,0.8797866503397623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,32,0.0032879999942249725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,5120,0.01902222302224901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,12288,0.02604977786540985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,768,0.02827111217710707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,512,0.0037919998996787598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,1024,0.0041866666740841335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,5120,0.02938488788074917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,51200,0.42324622472127277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,12288,0.043613334496816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,8192,0.04367022381888496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,8192,0.27197689480251735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,51200,0.11075822512308757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,768,0.01516355574131012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,1536,0.005859555469618903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,128,0.004253333227501975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,2048,0.05117688907517327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,6144,0.15923555692036948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,16384,10.52177005343967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,1536,0.04647555616166857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,7168,0.01963733302222358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,256,0.015506666567590503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,1024,0.00888000014755461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,3072,0.018082666728231642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,65536,11.263573540581596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,8192,0.023963555693626404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,32,0.00349866681628757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,7168,0.055594669448004834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,51200,0.17152978314293754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,768,0.007611555357774098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,12288,0.029311110575993855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,32,0.014145778285132514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,51200,0.05139288968510098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,3072,0.009546666509575313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,12288,0.5310364299350315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,2560,0.06347555584377713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,51200,6.45250744289822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,2048,0.010826666322019366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,32,0.039057777987586126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,4096,0.20896355311075845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,3584,0.16591644287109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,4096,0.399982213973999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,4096,0.02200444373819563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,12288,0.1823964383867052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,1536,0.24958133697509766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,5120,0.022482666704389784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,10240,0.03327288892534044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,128,0.0048142220411035745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,768,0.005958222266700532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,8192,3.4570551978217234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,256,0.007321777443091075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,7168,0.10842044485939874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,2560,0.008719111482302347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,12288,0.08319022258122762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,6144,0.026392888691690233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,128,0.005212444398138259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,6144,0.2638711134592692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,16384,0.1861306693818834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,12288,0.04811733298831516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,65536,0.08307200007968478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,256,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,3584,0.009311111436949836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,2048,0.006823110911581252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,256,0.004430222428507275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,256,0.006647999915811751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,6144,0.028671112325456407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,2048,0.006091555373536215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,10240,0.05757688813739353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,16384,2.604518254597982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,65536,0.015005333556069268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,3072,0.00941244430012173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,1536,0.021969777014520433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,1024,0.004059555629889171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,6144,0.024499555428822834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,7168,0.10109066963195801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,128,0.004050666673315896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,768,0.005693333430422678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,128,0.005179555465777715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,256,0.004552000098758274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,64,0.11412089400821263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,2560,0.13117867045932347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,4096,0.021913778450753953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,10240,0.030728889836205378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,65536,0.033632887734307185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,768,0.04116355710559421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,12288,0.3129884401957194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,64,0.005993777679072485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,16384,0.026516444153255884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,1536,0.009767110976907942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,1536,0.01958755486541324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,65536,0.13160178396436903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,6144,0.03298755486806234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,256,0.00872622181971868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,12288,0.010412444670995077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,2560,0.8565013143751355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,12288,0.030427555243174236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,768,0.02162222233083513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,512,0.004230222354332606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,2560,0.030748443471060857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,7168,0.016580444243219163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,10240,0.0424231125248803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,256,0.005423111220200856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,10240,0.7711066669887967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,8192,0.011668444507651858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,64,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,2048,0.04823644293679131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,128,0.005314666777849197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,512,0.006822222222884496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,51200,0.40546843740675187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,10240,0.49109956953260636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,2048,0.017090666625234816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,256,0.00453599997692638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,2560,0.00810222244924969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,128,0.02033422225051456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,65536,1.250165303548177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,1536,0.03533600105179681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,256,0.003125333330697484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,768,0.08362844255235459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,32,0.006333333336644703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,256,0.028449777099821303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,32,0.005688888745175467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,4096,0.017596445149845548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,16384,0.0359280010064443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,12288,0.030728889836205378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,8192,0.01723911033736335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,128,0.0027679999669392905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,128,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,12288,0.2340933216942681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,2560,0.00817599975400501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,7168,0.014217777384652032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,7168,0.07677066988415189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,3584,0.009215999808576372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,512,0.00646311127477222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,6144,0.02459377712673611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,51200,0.024489776955710515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,5120,0.039507554637061224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,2560,0.10247911347283258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,64,0.03621777892112732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,2048,0.02095288866096073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,64,0.0026213334252436957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,10240,0.20310844315422905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,6144,0.06524178054597643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,8192,0.413973331451416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,512,0.007119999991522894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,128,0.0058088890380329555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,65536,0.03874844312667847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,4096,5.5829039679633246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,51200,1.9321999020046656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,7168,2.990463044908312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,51200,0.20646933714548746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,3584,0.007691555553012424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,2560,0.057320886188083224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,1536,0.017283555534150865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,64,0.004820444517665439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,768,0.03927822245491876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,16384,0.8461706903245715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,768,0.04097955425580343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,10240,0.03703466720051236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,8192,0.00794488853878445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,3584,0.16882310973273384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,2048,0.39624799622429746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,3584,0.49336708916558164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,32,0.01592088904645708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,64,0.00408000002304713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,7168,0.02581866747803158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,3072,0.0395413339138031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,128,0.00999466660949919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,3072,0.020356444848908316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,3584,0.020046222541067336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,16384,1.28656800587972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,8192,0.33313510153028697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,12288,0.16053511036766901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,128,0.0032675556010670135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,2048,0.022208889325459797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,51200,0.24425954288906523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,65536,0.15886933273739284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,7168,0.006183111005359226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,51200,0.53020265367296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,2048,1.4553724924723308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,12288,0.028573334217071533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,32,0.003245333416594399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,3584,1.2100906372070312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,8192,0.015102222561836243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,51200,0.23853866259256998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,4096,0.01532799998919169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,6144,0.01132266637351778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,128,0.010118222071064843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,7168,0.013603555659453074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,5120,0.025825776987605627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,256,0.004865777575307422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,65536,0.49123467339409727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,3072,0.012465777496496836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,768,0.03606399893760681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,32,0.00776533368561003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,51200,0.30838576952616376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,2560,0.010069333016872406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,1024,0.05071555574735006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,3584,0.05530844463242424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,128,0.005043555465009478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,7168,0.05951466825273302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,7168,0.02219733264711168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,16384,0.35053155157301163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,64,0.010052444206343757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,5120,0.031375110149383545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,51200,0.11941511101192898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,2560,0.006624889042642381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,51200,61.92133246527778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,256,0.004387555436955558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,7168,0.6687857839796277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,65536,0.23015912373860678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,51200,0.061588446299235024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,2560,0.02977422210905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,1024,0.007188444336255391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,65536,2.699672910902235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,512,8192,0.015613334046469795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,65536,0.5908880233764648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,8192,0.014369777507252164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,6144,0.007191999918884701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,1024,0.026710222164789837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,10240,0.04341244366433886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,64,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,51200,0.13166044818030462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,12288,0.026408000124825373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,65536,0.1907351149453057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,10240,0.09371288617451985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,32,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,6144,0.010127999716334874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,1024,0.011866666376590729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,12288,0.31403199831644696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,7168,0.015896888242827523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,1024,0.004875555634498596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,768,0.004701333327425851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,51200,1.107466697692871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,6144,0.1210284498002794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,1024,0.011962666279739805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,7168,0.05428266525268555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,128,0.02648799949222141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,1536,0.010618666807810465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,4096,0.08251733250088163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,8192,0.36545599831475156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,12288,0.032383110788133405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,256,0.27906844351026744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,8192,0.0225600004196167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,5120,0.010788444843557147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,6144,2.819946712917752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,51200,0.19057688448164198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,51200,17.12090386284722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,8192,0.05613333649105496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,5120,0.006561777657932705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,32,0.007087111473083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,2560,0.020452444752057392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,10240,0.06655110915501912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,512,0.009276444713274637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,2048,0.2292684449089898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,5120,0.04567733407020569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,4096,0.00906133320596483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,128,0.006702222343948152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,16384,0.05885244740380181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,16384,0.3512702253129747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,3584,0.17011466291215685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,512,0.006077333456940121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,2560,0.012453333371215396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,65536,0.08721599976221721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,7168,0.0207004447778066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,768,0.007697777615653143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,768,0.019063111808564927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,3072,0.01884888940387302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,256,0.010108444425794814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,1536,0.00868088917599784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,1536,0.00675999994079272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,768,0.021612443857722815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,1024,0.03233333428700765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,3584,0.007832888927724626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,2560,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,32,0.005610666755172942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,2560,0.006950221955776215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,2560,0.008854221966531541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,768,0.2323404418097602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,65536,1.7190639707777235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,7168,0.8247049119737414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,128,0.005110222018427319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,6144,0.016493333710564505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,2048,0.041998220814598926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,65536,0.060580445660485156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,12288,0.5510302119784886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,1024,0.008386666576067606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,7168,0.019856888386938307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,3584,0.11603644159105088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,10240,0.3304284413655599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,3584,0.014848000473446317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,64,0.013304000099500021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,6144,0.011780444118711682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,8192,3.182679070366753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,12288,3.7045491536458335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,512,0.07660177681181166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,64,0.01427288850148519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,6144,0.008081778056091731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,4096,0.01020088874631458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,1024,0.00923288861910502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,4096,0.7720426453484429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,7168,0.03898577888806661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,32,0.0032239999208185407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,1536,0.0075635554061995606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,16384,0.04040088918473985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,768,0.004159111115667555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,512,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,512,0.006025777922736273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,2560,0.04902044600910611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,128,0.07984177933798896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,10240,0.018252443936136033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,65536,0.09154044257269965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,51200,0.06533688969082303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,51200,0.07039111190372042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,768,0.012680889003806643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,10240,0.01886755559179518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,16384,0.02049333353837331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,5120,0.018381334013409086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,8192,0.011881777809725868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,2560,0.008135110967689091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,4096,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,512,0.007643555601437886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,4096,0.03429244293106927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,128,0.08638311094707912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,768,0.022571555442280237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,3072,0.1344453361299303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,1024,0.00666222224632899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,51200,0.05347377724117703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,12288,0.3403093285030789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,2560,0.010669333239396414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,128,0.005589333259397083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,7168,0.1277351114484999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,2048,0.006481777876615524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,768,0.038133333126703896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,6144,0.019898666275872123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,256,0.004368888835112254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,2560,0.0075084442893664045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,256,0.006875555548402999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,128,0.005048000150256687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,32,0.02679555614789327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,8192,0.012299555871221753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,1024,0.01015644437736935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,1024,0.01166044423977534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,7168,0.05283911029497782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,3584,0.02500711050298479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,256,0.004090666770935059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,51200,0.23161421881781685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,128,0.004150222159094281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,32,0.015296889675988091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,1536,0.006256000035338932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,7168,0.024143111374643113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,10240,0.04880977670351664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,128,0.004119111018048393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,12288,7.655771891276042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,12288,0.008442666795518663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,7168,0.025566221939192876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,1536,0.0319199992550744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,32,0.00980711148844825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,32,0.029648890097935993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,51200,0.7824844254387749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,768,0.24907999568515352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,1536,0.08226133055157132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,2560,0.15678311718834773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,3072,0.1017013324631585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,3584,0.013003555436929068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,12288,0.6296462482876247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,8192,0.17813866668277314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,8192,0.012854222622182635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,256,0.04818488823042976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,768,0.006643555644485686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,12288,9.956143697102865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,6144,0.021915555000305176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,3584,0.01956888867749108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,4096,0.40874576568603516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,1024,0.1318302286995782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,512,0.014044443766276041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,6144,0.033935109774271645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,7168,0.1399137841330634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,768,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,65536,1.0143813027275932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,10240,0.03491288754675124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,7168,0.01846311158604092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,128,0.011122666299343109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,64,0.0026968889352348116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,8192,0.03477333320511712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,4096,0.32377423180474174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,7168,0.02759644389152527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,16384,0.01889600025282966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,4096,0.08487377564112346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,4096,0.3735342290666368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,768,0.3102204534742567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,2048,0.01127377814716763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,3584,0.022791999909612868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,256,0.00758133331934611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,3584,0.009272889130645329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,256,0.01717688971095615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,3072,0.13316533300611708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,512,0.0037946667936113146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,256,0.003195555466744635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,51200,0.08456889126035903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,768,0.006139555739031897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,128,0.002687999978661537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,32,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,256,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,128,0.0028275555620590844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,3072,0.008642666869693333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,2560,0.06975555419921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,32,0.0052488889131281115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,2048,0.1549875603781806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,768,0.006342222293217977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,1024,0.004250666747490565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,2560,0.05803022119734022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,7168,0.0236817780468199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,65536,0.06854933500289917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,7168,0.019724445210562814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,2560,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,4096,0.01982311076588101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,256,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,64,0.0028391112056043413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,3584,0.020861332615216572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,6144,0.016154666741689045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,2560,0.08374222119649251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,1024,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,512,0.005672889037264719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,2560,0.03273066547181871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,8192,0.218666672706604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,12288,0.21455555491977266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,10240,1.2902871237860787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,7168,0.01459822224246131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,256,0.15305777390797934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,65536,0.7137786547342936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,4096,0.6568453576829698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,1536,0.011223110887739392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,5120,0.01995466649532318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,3584,0.03687289026048448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,7168,0.01346577786737018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,3584,0.017902221944597032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,768,0.04073955615361532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,64,0.011880888707107969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,2048,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,7168,0.013448889056841532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,16384,0.2108515501022339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,1536,0.018722666634453666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,1536,0.008476444416575963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,512,0.0069448889957533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,3584,0.02197333342499203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,4096,0.039686222871144615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,256,0.004363555461168289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,512,0.006248888870080312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,256,0.03717422154214647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,6144,0.020424889193640817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,128,0.03251911203066508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,8192,0.009727111293209923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,51200,0.19179911083645293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,512,0.026692445079485577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,64,0.0025946667624844448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,4096,0.01091466678513421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,1536,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,2048,0.007340444458855524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,512,0.01408622165520986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,768,0.00526311124364535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,3072,0.009658666948477427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,3072,0.015492444237073263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,64,0.12215556038750543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,2048,0.09102577633327907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,10240,0.032560888263914324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,65536,0.13775822851392958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,10240,1.4585884941948786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,768,0.3438453409406874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,512,0.006648889018429651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,51200,0.35799466239081484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,3072,0.007623111208279927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,8192,0.02931555443339878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,6144,0.9429662492540148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,6144,0.10905244615342881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,6144,0.014497778481907315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,512,0.003660444584157732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,3584,0.02955289019478692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,12288,0.015847111741701763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,5120,0.15308266215854222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,12288,0.9009493721856011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,10240,0.03173866536882188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,32,0.010144888526863521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,12288,0.19414756033155653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,1536,1.2300444708930123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,8192,0.4374853240119086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,2560,0.011122666299343109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,51200,0.6524728669060601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,3072,0.15554132726457384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,4096,0.007901333272457123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,512,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,2560,0.11631288793351914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,7168,0.12008178234100342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,512,0.0038133333954546186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,512,0.00388711111413108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,128,0.005015111217896144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,16384,0.0240000006225374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,32,0.015255999233987598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,3584,0.02828177809715271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,128,0.0033946666452619764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,512,0.02183644473552704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,6144,0.013930666777822705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,10240,0.019176888797018263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,2560,0.006984000404675801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,7168,1.979329850938585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,10240,0.39721422725253636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,3584,0.014710222681363424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,7168,0.03499999973509047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,1024,0.030327998929553564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,512,0.0033777778347333274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,1024,0.037632889217800565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,6144,0.011604444848166572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,768,0.00702400008837382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,1536,0.008757332960764566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,6144,0.0704666641023424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,6144,0.20855732758839926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,1024,0.04342755675315857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,256,0.011808888779746162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,1536,0.01607466737429301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,16384,0.06271466943952772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,1024,0.030136889881557886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,3072,0.008886222210195329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,65536,2.4229297637939453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,256,0.01755466726091173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,128,0.02093066606256697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,2048,0.009080000221729279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,768,0.008779555559158325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,1536,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,12288,8192,0.05345066719584995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,5120,0.010069333016872406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,256,0.04440444376733568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,10240,2.0576008690728083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,512,0.005407111098368962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,1024,0.005223999834722943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,2560,0.07638311386108398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,768,0.008808888494968414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,128,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,32,0.029226667351192896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,128,0.021731555461883545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,3072,0.01553066737122006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,3584,1.67285950978597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,5120,0.01146488885084788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,10240,0.7307582431369357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,16384,0.08491288953357273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,10240,0.28298844231499565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,1536,0.028652442826165095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,5120,0.011409777734014722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,8192,0.14295910464392766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,512,0.09028355280558269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,10240,0.0703644421365526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,128,0.0047688889834615915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,256,0.005571555760171678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,51200,0.2648186683654785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,3584,0.041238221857282854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,32,0.02123911182085673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,128,0.003027555636233754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,256,0.06413422028223674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,32,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,2560,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,65536,0.34694578912523055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,51200,0.08335200283262466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,128,0.008858666651778752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,51200,0.05815022521548801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,2048,0.006726222319735422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,256,0.014881778094503613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,3584,0.045179555813471474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,1024,0.028470221492979262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,2048,0.006656888872385025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,3072,0.06683378087149726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,51200,0.021929777330822412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,65536,0.1903475522994995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,51200,0.09601333406236436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,2048,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,51200,0.38828176922268337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,4096,0.4789866871303982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,512,0.011964444484975604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,12288,0.3338826762305365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,16384,0.42398489846123594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,64,0.003079111170437601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,128,0.006105777704053455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,65536,0.39670221010843915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,5120,0.015106666419241162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,3072,0.016792888442675274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,65536,0.019304000669055514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,3584,0.030743112166722614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,7168,0.018409777018758986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,2048,0.010850666297806634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,65536,0.14743110868665907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,10240,0.02490577763981289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,2560,0.044093334012561373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,5120,0.6392364501953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,1024,0.01827022267712487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,3584,0.6305528746710883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,768,0.007407111426194509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,2048,0.00999111102686988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,64,0.003999999827808804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,5120,1.3566818237304688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,512,0.013821333646774292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,7168,0.09345510933134292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,128,0.03386044502258301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,4096,0.11087910996543036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,1536,0.015615999698638916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,128,0.0034764446318149567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,768,0.16737155119578043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,2048,0.01640266676743825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,1024,0.01184977756606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,6144,0.06957599851820204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,32,0.019187556372748483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,1024,0.012121777567598553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,3584,0.15729688273535833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,256,0.023738667368888855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,3072,0.05652711126539442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,32,0.005049777941571342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,5120,0.08448977602852716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,4096,0.03408533334732056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,4096,0.012057777908113269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,51200,4.835068596733941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,128,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,5120,0.009248889154858058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,10240,0.3715413411458333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,12288,0.017092444830470614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,1024,0.006037333359320958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,3584,0.07153866688410442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,65536,1.0891218185424805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,5120,0.02091288897726271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,12288,0.01421333352724711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,1024,0.007621333003044128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,16384,0.024466666910383437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,6144,0.010125333236323463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,4096,0.006129777679840724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,512,0.008261333737108443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,768,0.0038088887102074097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,768,0.0037671112351947357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,7168,1.3703724543253581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,10240,0.017295999659432303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,768,0.01718222267097897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,3584,0.03344444433848063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,3072,0.09577688905927871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,12288,0.03914755582809448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,768,0.003967111309369405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,256,0.0037662221325768363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,7168,0.1259688933690389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,10240,0.007355555891990662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,128,0.006752000086837345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,2048,0.012284444438086616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,16384,0.05178755521774292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,12288,0.04272355635960897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,7168,0.027035555905765955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,128,0.004342222379313575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,256,0.003031999907559819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,10240,0.029083556599087183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,16384,0.08956355518764919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,128,0.0041848888827694785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,65536,0.08199110958311293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,51200,0.2742435667249891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,1024,0.0136817776494556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,512,0.03559199968973795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,64,0.016535111599498324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,2048,0.09720089038213094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,12288,0.015626667274369132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,32,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,512,0.004460444466935264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,1024,0.07179466883341472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,512,0.004369777937730153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,2048,0.030801776382658217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,8192,0.026659554905361597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,6144,0.1159848901960585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,4096,0.13942044311099583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,2048,0.7573875321282281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,768,0.02008888953261905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,3584,0.00944977750380834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,51200,0.23087112108866373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,7168,0.012488000094890594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,2560,0.014935111006100973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,51200,0.37146133846706814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,2048,0.2482995457119412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,32,0.013012444807423485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,5120,0.13867466979556614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,7168,0.01260355528857973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,8192,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,1024,0.006924444602595435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,4096,0.13806845082177058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,256,0.004751111070315043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,768,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,12288,0.044617778725094266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,768,0.01791288952032725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,2048,0.08401688602235581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,3072,0.020655110478401184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,4096,0.03084711233774821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,2048,0.03997955719629923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,768,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,64,0.12783911493089464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,64,0.01609244445959727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,6144,0.11339733335706924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,7168,0.18840621577368843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,2048,0.008107555409272512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,4096,0.007607999775144789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,256,0.11644711759355332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,128,0.1131955517662896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,256,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,7168,0.014431110686726041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,7168,0.06585422489378187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,12288,0.025935111774338618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,2048,0.020503110355801053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,64,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,4096,0.0136817776494556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,7168,0.05643466446134779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,5120,0.08506488800048828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,2560,0.007792889244026608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,7168,0.08940444389979045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,12288,4.295868343777127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,1024,0.008991999758614434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,5120,0.04244177871280246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,3584,0.020399111840460036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,768,1.5507013532850478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,64,0.0027520000520679685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,1536,0.006126222097211414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,512,0.0068853331936730286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,7168,0.01739911072784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,6144,0.2021733389960395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,1536,0.018713777263959248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,64,0.002567111204067866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,768,0.008186666501892937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,32,0.04589955674277412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,16384,0.3145279884338379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,51200,0.43292445606655544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,10240,0.019208888212839764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,3584,0.010959111154079437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,32,0.010634666515721215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,10240,0.019329778022236295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,3584,0.13104800383249918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,128,0.005151999907361137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,5120,0.2695591184828016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,8192,0.022336888644430373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,12288,1.447580443488227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,1024,0.010652444428867765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,1536,0.009617778162161509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,1536,0.034244444635179304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,51200,1.3486969206068251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,5120,0.11720621585845947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,8192,0.014898666077189975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,10240,3.0576169755723743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,16384,0.06256355841954549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,16384,0.08806488911310832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,64,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,5120,0.043324444029066295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,7168,0.02567377852069007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,512,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,10240,0.04244177871280246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,12288,0.1742711067199707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,768,0.008159110943476358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,1024,0.006112888869312074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,16384,0.04370933439996508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,16384,0.809791088104248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,256,0.005734222216738596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,2048,0.05394311083687676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,10240,0.018548443913459778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,1536,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,6144,0.016727111405796476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,3584,0.009832888841629028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,6144,0.4018639988369412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,256,0.0034204444123639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,5120,0.16596444447835287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,2560,0.013054221868515015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,32,0.036807109912236534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,65536,18.501051161024304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,6144,0.041328890456093684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,10240,0.02678488857216305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,768,0.40511377652486164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,5120,0.008100444244013892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,512,0.02223199937078688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,768,0.004275555411974589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,5120,0.006908444480763541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,256,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,8192,0.2326408757103814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,512,0.06722399923536512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,16384,0.021922666165563796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,256,0.0032213332338465583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,3072,0.13768088817596436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,64,0.021159110797776118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,65536,0.06920088662041558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,10240,0.007343110938866933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,7168,0.31287556224399143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,768,0.021055999729368422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,4096,0.013461333182122974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,128,0.0029582221888833572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,4096,0.008316444853941599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,1536,0.027216888136333887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,1536,0.051853332254621715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,65536,0.40184712409973145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,5120,0.01349955548842748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,1024,0.04126755396525065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,32,0.00657066661450598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,32,0.0029475556479560006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,2560,0.018947554959191214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,512,0.021014221840434607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,65536,0.24560888608296713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,8192,0.040516446034113564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,7168,0.05184444453981188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,8192,0.053885334067874484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,16384,0.8399128913879395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,12288,0.03997155692842271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,10240,0.06309600008858575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,4096,0.26021332210964626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,2560,0.009583999713261923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,7168,0.017878222796652053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,3584,0.012938666674825879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,5120,0.2755626572502984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,5120,0.05022222134802076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,5120,1.0143804550170898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,51200,0.24506576855977377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,768,0.018259555101394653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,2560,0.08992444144354926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,512,0.005027555757098728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,16384,0.15846221976810032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,32,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,3072,0.044030222627851695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,3072,0.008582221964995066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,1024,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,7168,0.024873778223991394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,32,0.012060444388124677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,8192,0.037596444288889565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,2048,0.09089244736565484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,10240,0.054823110500971474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,4096,0.01090844472249349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,5120,0.015446222490734525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,768,0.005977777971161737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,51200,0.06466844346788195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,1024,0.00444444434510337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,256,0.20060622692108154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,6144,0.04725777771737841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,3584,0.007717333734035492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,16384,0.12969422340393066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,128,0.004332444320122401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,2560,0.012513777448071374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,12288,0.047761778036753334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,128,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,51200,16.823538886176216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,768,0.007956444389290279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,3072,0.12116977903578018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,1024,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,7168,0.05447555581728617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,3584,0.011154666543006897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,512,0.01072799993885888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,8192,0.015834665960735746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,2048,0.08335822158389622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,7168,0.054735110865698926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,10240,0.09213244252734715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,5120,0.03174044357405768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,3072,0.3909662299686008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,3072,0.7895190980699328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,256,0.013663999736309052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,768,0.009528000321653154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,12288,0.49639733632405597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,7168,0.28687911563449436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,65536,1.9846977657741969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,16384,0.03597955571280585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,7168,0.02907466557290819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,10240,0.027423999375767175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,7168,0.05125777920087179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,2560,0.010579555398888059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,12288,0.02554399934079912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,1536,0.007425777614116669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,768,0.0057813334796163775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,3072,0.3115777704450819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,2560,0.041492445601357356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,2560,0.08731644683414036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,65536,0.07063733206854926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,1024,0.019152889649073284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,7168,0.22478577825758192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,1536,0.04826133449872335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,51200,0.13790222009023032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,1024,0.005767999837795894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,64,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,16384,0.06586933135986328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,32,0.007164444360468123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,16384,0.026629333694775898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,5120,0.5645973417494032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,8192,0.01368444412946701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,2048,0.016374222106403775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,2560,0.04203644394874573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,2560,0.024336000283559162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,16384,0.045816000964906484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,1536,0.006063111126422882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,2560,0.032319999403423734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,1024,0.013162666724787818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,6144,0.03659644391801622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,10240,0.3765386740366618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,5120,0.02583644456333584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,5120,0.03523555397987366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,2048,0.05130666825506422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,2048,0.3228648768530952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,10240,0.01725955473052131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,3072,0.0158933334880405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,7168,0.0718053314420912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,2560,0.009485333330101436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,2560,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,1024,0.022525333695941504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,65536,0.14046311378479004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,6144,0.2697226736280653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,64,0.005130666825506422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,12288,0.029507554239696924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,3072,0.015356444650226168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,3584,0.00793599999613232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,7168,0.040779554181628756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,2048,0.005995555470387141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,512,0.00328533330725299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,64,0.00639111093348927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,128,0.0032257777121331957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,12288,0.30810311105516225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,128,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,1024,0.006455111006895701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,5120,0.09856533341937596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,6144,0.3367137908935547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,7168,0.08948177761501735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,512,0.011459555890825061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,512,0.0057946667075157166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,6144,0.1378648943371243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,10240,0.05213155680232578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,3584,0.014905777242448596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,10240,0.009662222531106737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,5120,0.16017511155870226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,5120,0.03491466575198703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,512,0.0037751110891501107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,2048,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,51200,0.8278631104363335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,1024,0.03582133187188043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,3072,0.014561777313550314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,65536,0.033349331882264875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,128,0.00816533300611708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,51200,0.05273600088225471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,3584,0.07605599694781832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,3584,0.02736977736155192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,65536,0.1888577805625068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,768,0.03193333413865831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,2048,0.010694222317801582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,6144,0.018024888303544786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,16384,0.6950186623467339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,10240,0.030360887447992962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,32,0.040567109982172646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,2048,0.012349333200189801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,3584,0.023399111297395494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,3072,0.029402666621738013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,2560,0.007157333195209503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,1536,0.006832888970772426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,256,0.006416888700591193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,1536,0.007684444387753804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,3584,0.055013332102033824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,8192,0.01514755520555708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,2048,0.030444443225860596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,51200,0.19233422809176973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,6144,0.034105777740478516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,6144,0.0262773334980011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,128,0.0041315555572509766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,65536,0.633662223815918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,16384,0.6091661983066136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,1536,0.09848533074061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,10240,0.11662399768829346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,3584,1.5746160083346898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,8192,0.024125332633654278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,10240,0.11279200183020698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,7168,0.012874667015340594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,16384,0.0444124440352122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,16384,0.058633777830335826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,768,0.009110222260157267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,16384,0.5510471132066515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,5120,0.07840000258551703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,2048,0.006958222223652734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,1024,0.01037599974208408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,512,0.008784888519181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,3584,0.007483555210961237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,128,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,3584,0.020463999774720933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,16384,0.04520977867974175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,65536,0.4932417869567871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,512,0.009762666291660732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,16384,0.1946248875723945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,6144,0.03631466627120972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,1536,0.07129333416620891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,2560,0.017665778597195942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,256,0.0053324442770746015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,8192,0.4848968717787001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,3072,0.2468035486009386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,3584,0.00903733323017756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,5120,0.7928613556755914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,8192,0.03752355443106757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,4096,0.06768533256318834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,12288,0.4974062177870009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,128,0.005120888766315249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,1536,0.16146932707892522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,2048,0.007455999652544658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,768,0.005560889012283749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,8192,0.027071999178992376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,2048,0.04153155618243747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,3584,0.012661333713266583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,1536,0.0107022225856781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,3072,0.008052444292439355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,12288,0.08841600020726521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,5120,0.11748354964786106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,256,0.07051022185219659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,51200,1.3809866375393336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,1024,0.5217715369330512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,256,0.00692622239391009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,512,0.033992889854643084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,5120,0.006239999913507038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,7168,0.014492443866199918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,51200,0.29990487628512913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,1024,0.004975111120276981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,1536,0.01423822177780999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,8192,0.2289644347296821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,5120,0.041251556740866765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,4096,0.0059164443777667145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,10240,0.023452444208992854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,3072,0.0251875552866194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,12288,0.04025600022739834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,256,0.1265137725406223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,512,0.018398221996095445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,6144,0.02647999922434489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,2560,0.050140443775388926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,1536,0.02923999892340766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,12288,0.017697778013017442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,51200,0.06379199690288968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,128,0.01961599952644772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,2048,0.007090667055712805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,256,0.024307555622524683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,65536,0.7330942153930664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,3584,0.017335999343130324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,4096,0.02970400121476915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,51200,13.56840345594618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,8192,0.03516177667511834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,3072,0.02425688836309645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,768,0.014489778214030795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,2560,0.011172444456153445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,8192,0.022403554783927068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,2048,0.007135111424658034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,4096,0.010586666564146677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,2560,0.006474666711356904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,32,0.015118221441904703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,51200,0.11664266056484646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,6144,0.01964977714750502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,4096,0.08021066586176555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,12288,0.09896177715725368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,1536,0.006271111054552927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,7168,0.12649510966406927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,1024,0.003726222034957674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,12288,0.014626666903495789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,8192,0.33980533811781144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,7168,0.06654489040374756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,6144,0.01368444412946701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,3584,0.018791110979186166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,16384,0.12589155303107366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,7168,0.1840640041563246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,3072,1.4353218078613281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,32,0.006597333484225803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,512,0.009081777599122789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,5120,0.010693333215183683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,7168,0.9639333089192709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,7168,0.01866044435236189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,3072,0.03484800126817491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,6144,0.15665245056152344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,512,0.017492444978819955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,64,0.030709332889980737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,512,0.005535111245181825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,4096,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,512,0.00562755556570159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,2048,0.012153777811262341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,512,0.004430222428507275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,3584,0.026929777529504564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,2560,0.008678221868144141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,6144,0.9426604376898872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,10240,0.014535999960369535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,4096,0.07596533166037665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,6144,0.019186667270130582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,7168,0.030132446024152968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,256,0.03333688775698344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,51200,0.018012444178263348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,3072,0.14813421832190618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,128,0.017117333081033494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,16384,0.10237244764963786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,8192,2.1478790707058377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,10240,0.1505875587463379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,3584,0.028396444188223943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,128,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,1536,0.05765777826309204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,6144,0.007065777977307637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,5120,0.32351554764641655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,3584,0.008479110896587372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,512,0.009438222481144799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,3072,0.04675288995107015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,64,0.005327111317051782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,2560,0.14428622192806667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,51200,0.01294133315483729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,2560,0.00942311104800966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,8192,0.027280888623661462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,256,0.00721155520942476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,128,0.005173333403136996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,512,0.00611200018061532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,32,0.0027706666539112725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,65536,0.18699288368225098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,3072,0.022542221678627863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,51200,0.10607200198703343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,3072,0.13640711042616102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,1024,0.17365154955122206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,128,0.017477333545684814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,7168,0.00657066661450598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,2560,0.0064071110553211635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,768,0.022543110781245764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,6144,0.023379556006855432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,16384,5.972731696234809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,6144,0.26324354277716744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,5120,0.013643555343151093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,12288,0.21945599714914957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,6144,0.049248887432946094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,6144,0.10702488819758098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,256,0.004402666870090696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,3072,0.016920000314712524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,7168,0.013384888569513956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,2048,0.046441776884926685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,256,0.020265777905782063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,7168,0.035587555832333036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,768,0.009958222508430481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,65536,0.1626488897535536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,65536,0.5274977684020996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,12288,0.026856889327367146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,4096,0.0207004447778066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,1024,0.5331217977735732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,2560,0.007277333074145847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,12288,0.056672890981038414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,512,0.007126222054163615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,8192,0.014720888601409065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,2048,0.055474665429857045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,10240,0.014800000521871777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,768,0.07089866532219781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,4096,0.23475644323560926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,256,0.005576888720194499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,128,0.00518844442235099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,768,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,768,0.013008000122176277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,768,0.21403643820020887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,8192,0.35060445467631024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,512,0.014741332994567024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,2048,0.041645334826575384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,64,0.010513777534166971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,3584,0.7060737609863281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,64,0.0032382222513357797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,12288,0.025569778349664476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,256,0.21709067291683623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,768,0.04507199923197428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,32,0.008791999684439765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,8192,0.1005057758755154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,512,0.023249778482649062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,16384,0.35680532455444336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,2048,0.02167822255028619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,256,0.015598222613334656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,5120,0.017816000514560275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,65536,0.11865599950154622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,256,0.013122667041089801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,12288,0.14936178260379368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,16384,0.01850755512714386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,512,0.006217777729034424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,65536,0.07736888859007093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,2048,0.9603332943386502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,768,0.005331555588377847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,6144,0.033281776640150286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,1024,0.009672889278994666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,1024,0.010100444157918295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,5120,0.08498311042785645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,1024,0.008147555920812819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,3584,0.007268444531493717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,10240,0.07743910948435466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,128,0.005097777893145879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,1536,0.014752000570297241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,4096,0.019889778561062284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,3584,0.016545777519543965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,256,0.015295111470752291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,8192,0.015547555353906421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,1536,0.048010667165120445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,51200,2.0360808902316623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,768,0.006522666662931442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,4096,0.06772266493903266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,3584,0.13917777273390028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,6144,0.0325271122985416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,32,0.0031066667288541794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,16384,0.02371911042266422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,6144,0.01367911116944419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,128,0.06305155489179823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,768,0.005776888794369168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,2048,0.013642667068375481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,8192,0.0584391090604994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,8192,0.03429866830507914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,4096,0.009340444372759925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,1024,0.12680444452497694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,2560,0.22387822469075522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,512,0.00611200018061532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,12288,0.3983137872483995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,10240,0.271616882748074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,4096,0.03083289000723097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,1536,0.011220444407727985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,1536,0.023017777336968318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,32,0.009402666654851701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,65536,0.9076924853854709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,8192,0.06183911032146878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,10240,8.441111246744791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,51200,0.10108444425794814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,3584,0.1695457829369439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,10240,0.19733956125047472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,10240,0.11792977650960286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,16384,0.2165751059850057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,8192,0.011798222031858234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,1536,0.0073724447025193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,5120,0.021831111775504217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,65536,5.724097781711155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,512,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,3072,0.11529421806335449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,256,0.028146667612923518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,2560,0.03156266609827677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,16384,0.05813866853713989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,128,0.03796622157096863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,4096,0.019323555959595572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1536,51200,0.4416195551554362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,256,0.01020088874631458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,10240,0.0943288869327969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,3072,0.03208977646297879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,256,0.021088000800874498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,65536,0.10343733098771836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,2048,0.018271999226676095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,16384,0.2802026536729601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,6144,0.37265422609117294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,128,0.009272000028027428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,7168,0.14821155865987143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,64,0.004115555435419083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,10240,0.05097511079576281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,2560,0.015073777900801765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,3584,0.06236266427569919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,8192,0.6716906759474012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,3584,0.03395822313096788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,3584,0.3623937765757243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,7168,0.2603093253241645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,2048,0.16840622160169813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,1024,0.02288888891537984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,4096,0.03955022162861294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,3072,0.06678577926423815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,32,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,128,0.0047386665311124595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,1024,0.006896889044178857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,3072,0.04988266693221199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,1024,0.040623999304241605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,128,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,32,0.002567111204067866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,2048,0.007083555890454187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,512,0.02996444371011522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,768,0.006493333313200209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,8192,2.2349234686957464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,4096,0.023893333143658105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,16384,0.055608888467152916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,51200,2.7263387044270835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,2048,0.029362665282355413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,4096,0.015560888581805758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,2048,0.010288000106811523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,65536,0.1393431160185072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,128,0.003129777808984121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,768,0.023569777607917786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,3584,0.014893333117167154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,1024,0.006665777828958299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,3072,0.015328889091809591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,6144,0.0207306659883923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,5120,0.2308195432027181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,256,0.01163999984661738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,2560,0.10922933287090725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,5120,0.1657386620839437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,12288,0.04563911093605889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,51200,0.6290702290005153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,6144,0.05468622181150648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,2048,0.019946666227446664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,1536,0.008722667064931657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,3072,0.009329777624871995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,512,0.33276179101732045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,6144,0.011339555184046427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,7168,0.00903555585278405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,512,0.020522667302025687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,2048,0.00923466682434082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,2560,0.2090160051981608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,51200,0.5008533265855577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,12288,0.10410844617419773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,7168,0.08310133218765259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,4096,0.020455999506844413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,4096,0.02648977769745721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,2560,0.014308444327778287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,3072,0.02276355524857839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,2048,0.011591110792424945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,256,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,16384,0.02286044425434536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,1024,0.0037724446091387006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,256,0.012099555797047086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,1536,0.021495999561415777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,12288,0.017714665995703805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,1024,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,12288,0.24420976638793945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,65536,0.06559822294447157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,65536,0.06797511047787137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,65536,0.6509013175964355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,768,0.014119999276267158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,16384,0.24279732174343535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,64,0.006509333435032103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,256,0.022760000493791368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,4096,0.02844888965288798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,128,0.03427555494838291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,8192,0.019002666076024372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,64,0.0032888888898822996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,51200,7.571808709038629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,64,0.010671999719407825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,32,0.011180444724029966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,6144,0.04458399944835239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,512,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,2560,0.013819555441538492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,64,0.011513777905040316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,10240,0.019186667270130582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,128,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,1536,0.00813155538505978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,1024,0.02719377809100681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,65536,0.25778312153286403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,128,0.005437333136796951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,3584,0.057796445157792836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,65536,0.0690773328145345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,2048,0.006312888943486744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,51200,0.04974844389491611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,2560,0.02053777707947625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,12288,3.32460085550944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,3072,0.10264533095889622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,8192,0.031317333380381264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,256,0.013353777428468069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,2560,0.02000266644689772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,16384,0.04285244478119744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,3072,0.11637511518266465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,12288,0.020997333857748244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,3072,0.17249155044555664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,4096,0.20496088928646514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,768,0.011337777806652917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,7168,4.72499042087131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,10240,0.22024267249637178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,3584,0.014729777971903482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,12288,0.40891112221611875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,7168,0.049272888236575656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,4096,0.03212799959712558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,1024,0.03700444433424208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,1024,0.004230222354332606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,16384,0.022287999590237934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,2048,0.0181040002240075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,3072,0.2602275477515327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,16384,0.03978133201599121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,2048,0.026597332623269822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,3072,0.015134221977657743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,128,0.005987555616431766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,512,0.05246577660242716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,512,0.14370400375790066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,16384,0.018612444400787354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,1024,0.007515555454625024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,32,0.008567999634477828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,8192,0.08638222350014581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,32,0.24802843729654947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,2048,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,512,0.009195555415418413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,7168,0.04127466678619385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,65536,0.06544444296095106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,1024,0.018047110901938546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,5120,0.0473582214779324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,7168,0.05119199885262383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,32,0.0027439999911520216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,7168,0.12148000134362115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,5120,0.5397040049235026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,8192,0.0165928883685006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,65536,7.588236490885417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,512,0.009948444035318162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,12288,0.007521777517265744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,5120,0.33423911200629336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,4096,0.13010756174723306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,2560,0.006687110910813014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,6144,0.0325893329249488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,1024,0.06350933180914985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,32,0.027992000182469685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,12288,0.03252977795071072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,7168,0.11124177773793538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,2048,0.008255110846625434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,2048,0.16654311286078557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,128,0.0075635554061995606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,6144,0.22930399576822916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,8192,0.32639288902282715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,12288,0.048467556635538735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,1536,0.03221599923239814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,65536,0.1499697764714559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,64,0.006160000132189856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,10240,0.017252445220947266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,6144,0.4051537778642442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,7168,0.024784000383483037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,2560,0.13611822658114964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,8192,0.0166320006052653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,3584,0.04137333234151205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,128,0.017123555143674214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,3584,0.05892178085115221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,7168,0.0063822223908371395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,256,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,7168,0.010728889041476779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,512,0.006878222028414409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,8192,0.12214933501349555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,10240,2.7960018581814237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,65536,0.20891643895043266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,5120,0.01032266683048672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,4096,0.009861333502663506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,512,0.030156443516413372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,32,0.005893333504597346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,12288,0.15356355243259007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,2048,0.00723555518521203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,65536,0.3651288880242242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,7168,0.0196782218085395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,7168,0.030127998855378892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,3072,0.2682817776997884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,6144,0.010859555668301053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,7168,0.49079375796847874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,3584,0.009140444298585257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,65536,7.3851064046223955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,3584,0.017626666360431247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,65536,5.565798017713759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,7168,0.6012399991353353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,1536,0.011403555671374003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,10240,0.027395556370417278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,16384,0.038318223423428006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,7168,0.23667555385165742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,768,0.007931555310885111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,3072,0.09955555862850612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,16384,1.1385182274712455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,64,0.11488178041246201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,2048,0.007248889240953658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,768,0.008264888491895463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,65536,0.08345866865581936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,512,0.004649777793222004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,512,0.003997333347797394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,3584,0.011882666912343768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,7168,0.25177955627441406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,128,0.004918222212129169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,12288,0.03147644466824002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,256,0.005671111245950063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,10240,0.028055999014112685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,7168,0.012014221813943652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,4096,0.3407448927561442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,128,0.027659555276234944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,6144,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,4096,0.027797334724002417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,2048,0.048604445325003735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,256,0.00774133370982276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,6144,9.905815124511719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,256,0.01076533314254549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,6144,2.1815128326416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,51200,1.242018699645996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,128,0.004549333204825719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,3584,0.007721777591440413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,768,0.005613333235184352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,512,0.06119733386569553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,3072,0.016120000018013846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,32,0.002982222164670626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,1536,0.0070062221752272705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,4096,0.007752888732486301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,256,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,7168,0.48553779390123153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,2048,0.08557689189910889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,2560,0.011021333436171213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,3072,0.012692444854312472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,3584,0.03185155656602647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,128,0.10592622227138943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,768,0.006728000111050076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,3584,0.01535377734237247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,8192,0.0703279972076416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,2048,0.29669067594740123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,12288,0.04454400142033895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,2560,0.01676177812947167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,2560,0.013038222160604266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,256,0.004464000049564573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,1024,0.021433777279324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,51200,0.24082843462626138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,1024,0.00960622231165568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,5120,0.00941066692272822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,4096,0.020391111572583515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,3072,0.11330578062269424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,7168,0.6917262077331543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,64,0.004056888735956616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,6144,0.3449510998196072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,6144,0.029616889026429918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,512,0.003216000066863166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,2560,0.007463110817803278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,2560,0.08054755793677436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,7168,0.02088444431622823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,12288,0.03778666588995192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,7168,0.09853688875834148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,7168,0.12315377924177383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,65536,0.08417510986328125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,2048,0.006752000086837345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,3072,0.012236444486512078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,2560,0.015507555670208402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,10240,0.3844204478793674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,16384,1.0475386513604057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,5120,0.17537066671583387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,3072,0.11491910616556804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,128,0.011992000043392181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,16384,0.21528622839185926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,256,0.004518222063779831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,8192,0.012344000240166983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,8192,0.011750222080283694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,3072,0.009909333454238044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,128,0.18894844584994844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,51200,0.1694613297780355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,128,0.11218399471706814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,8192,0.013937777943081327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,12288,0.015408888459205627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,3072,0.008808888494968414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,2560,0.05108266737726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,768,0.004487110922733943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,8192,0.9748800065782336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,16384,0.11700622240702312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,10240,0.03545511100027297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,64,0.013458666702111563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,5120,0.042416890462239586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,8192,0.0373946660094791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,256,0.0037226668662495087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,128,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,2560,0.007521777517265744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,10240,0.2848195499844021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,6144,0.040785776244269475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,1536,0.011979555918110741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,3584,0.463755554623074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,6144,0.4228382110595703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,3072,0.021373333202468023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,32,0.034244444635179304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,2048,0.019367999500698514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,51200,2.2164230346679688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,10240,1.0027057859632704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,128,0.010223999619483948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,2048,0.21212266551123726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,16384,9.83920627170139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,1536,0.012294222083356647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,3072,0.06092622545030382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,7168,0.12089510758717854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,5120,0.022602667411168415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,4096,0.011866666376590729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,128,0.010237333675225576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,2560,0.12745600276523167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,2048,0.019205333458052743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,3584,1.7883182101779516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,512,0.0068808889223469635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,2560,0.026553778184784785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,32,0.0026986667265494666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,32,0.002528000002106031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,6144,0.006233777850866318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,768,0.009610666996902889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,16384,1.1052897771199544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,128,0.0042053332759274375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,51200,0.5825022061665853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,51200,0.2631964418623182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,16384,11.393504672580294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,512,0.06263911061816745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,2560,0.03527199890878465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,64,0.03351377778583103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,256,0.004976888911591636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,1536,0.0520977775255839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,1536,0.010047111246320937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,7168,0.01703288820054796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,128,0.006968888971540663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,2560,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,51200,0.06031911240683662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,5120,0.00674044465025266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,256,0.011905777785513135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,2048,0.06028799878226387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,64,0.007260444263617198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,51200,0.09515644444359674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,2560,0.01869244376818339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,3072,0.007495111061467065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,1536,0.30543822712368435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,7168,0.017987555927700467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,65536,0.055049777030944824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,65536,0.11246844132741292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,64,0.013046222428480784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,16384,8.422130160861546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,6144,0.4149484369489882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,1536,0.008374222450786168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,768,0.11682400438520645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,768,0.005056888692908817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,10240,0.055941330062018506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,2048,0.014298665854665967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,128,0.004994666824738185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,1024,0.027871999475691054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,3584,0.020996444755130347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,1024,0.006189333481921091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,16384,0.28203021155463326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,51200,0.07022044393751356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,3072,0.06301955382029216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,65536,3.9278479682074656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,1024,0.009993777506881291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,65536,0.2089386648601956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,4096,0.015007111761305066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,6144,0.019440889358520508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,768,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,65536,1.0715369118584526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,6144,0.014613333675596448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,6144,0.02586577832698822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,64,0.014977777997652689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,10240,0.010919999745157031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,8192,0.01532977819442749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,64,0.008428444465001425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,32,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,2048,0.0403404434521993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,16384,0.0372613337304857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,256,0.007281777759393056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,7168,0.020470221837361652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,3584,0.22439021534389922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,1024,0.00397866674595409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,2048,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,64,0.012997333374288348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,16384,0.02113688819938236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,65536,0.649361769358317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,12288,0.06276800235112508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,6144,0.010694222317801582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,7168,0.037286221981048584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,1536,0.03974577784538269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,768,0.009364444348547194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,1024,0.010784889260927836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,768,0.09895910819371541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,768,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,2560,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,12288,0.07169600327809651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,6144,0.17131643825107148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,512,0.009517333573765224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,3584,0.0199297782447603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,128,0.0047324444684717385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,1024,0.029676444000667993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,6144,0.15488978226979574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,2048,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,6144,0.011615999870830111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,4096,0.019367999500698514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,3584,0.07596177524990506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,1536,0.009511111511124505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,256,0.00589155571328269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,128,0.02908888790342543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,65536,0.009319999979601966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,32,0.002936888900068071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,512,0.005865777946180767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,1536,0.425294213824802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,6144,0.12707911597357854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,6144,1.9634710947672527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,1024,0.152911106745402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,8192,0.02280355493227641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,3072,0.02197333342499203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,16384,0.19747466511196562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,6144,0.1881164444817437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,2560,0.23012799686855742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,8192,0.06165066692564222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,256,0.00563733321097162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,8192,1.3604319890340169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,8192,0.08971822261810303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,16384,0.4356702168782552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,51200,4.092365264892578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,2560,0.015262222952312894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,128,0.011490667031870948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,10240,0.26304978794521755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,4096,0.010334221853150262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,4096,0.009024889104896123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,65536,0.16040889422098795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,10240,0.04073866539531284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,65536,0.06994133525424533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,64,0.0028693332440323303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,65536,0.7695973184373645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,12288,0.10869777864880031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,512,0.004039111236731212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,128,0.10326400068071152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,64,0.0069066666894488865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,1024,0.023381332556406658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,6144,0.03651288814014859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,2048,0.01144177797767851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,2048,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,65536,1.1210533777872722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,10240,2.3956932491726346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,5120,0.013608888619475894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,768,0.0038577777643998465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,65536,0.06264355447557238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,32,0.0052639999323421055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,51200,0.08636889192793105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,16384,0.2615955670674642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,3584,0.4050995508829753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,12288,0.023444443941116333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,1536,0.02858755654758877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,2048,0.009372444616423713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,7168,0.0269768883784612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,3072,0.18437688880496553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,6144,0.019948444432682462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,5120,0.010425777898894416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,512,0.00979022185007731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,8192,0.028247998820410833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,3072,0.0063031112982167145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,1536,0.006728888799746831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,2560,0.03491111265288459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,16384,0.42821333143446183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,7168,0.15144088533189562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,5120,0.03950400153795878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,6144,0.1086568898624844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,7168,0.2746524545881483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,64,0.02414488957987891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,65536,0.9622728559705945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,32,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,1024,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,8192,0.5802657869127061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,3072,0.01614755557643043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,2048,0.17667467064327666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,64,0.1955333285861545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,6144,0.02496799992190467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,3072,0.059086223443349205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,2048,0.2831253210703532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,32,0.01165066659450531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,8192,0.03125244379043579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,4096,0.3617768817477756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,8192,0.8269484308030871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,12288,0.13359288374582926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,256,0.011561777856614856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,1024,0.00628355559375551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,2048,0.007797333101431529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,1024,0.014501333236694336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,1024,0.021495110458797876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,512,0.012248000337017907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,3584,0.02222577730814616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,7168,4096,0.021801778011851843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,10240,1.5892204708523219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,3584,0.010447111394670276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,3072,0.03102133340305752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,2560,0.03711466656790839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,4096,0.08472088972727458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,6144,0.01440533333354526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,3072,0.01350311107105679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,51200,0.284313784705268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,1536,0.006303999986913469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,6144,2.1297083960639105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,7168,0.02927288744184706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,2048,0.04970488945643107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,8192,0.32377510600619847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,2048,0.08382755517959595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,256,0.002895111011134254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,4096,0.02407466702991062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,3584,0.010992888775136737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,65536,3.4493963453504772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,65536,0.00960888879166709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,8192,0.31254132588704425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,256,0.005864000154866113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,12288,4.828579372829862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,3584,0.015089778436554803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,7168,0.20760799778832328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,65536,0.5432897673712836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,3072,0.0151235560576121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,3072,0.0069324444565508105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,4096,0.009388444324334463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,12288,0.1754737827512953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,10240,0.2585671212938097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,2048,0.09077511231104533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,4096,0.023822221491071913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,51200,0.2180622286266751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,1024,0.061252441671159535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,2048,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,3584,0.019380443625979953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,8192,0.022392888863881428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,1024,0.0038000001675552795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,2560,0.04196088843875461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,3072,0.007081777685218387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,7168,0.03210044569439358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,5120,0.03236533204714457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,6144,0.016238222519556682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,64,0.002728888971938027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,4096,0.010932444698280759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,768,0.036178665028678045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,7168,0.03660622239112854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,128,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,128,0.12686933411492243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,10240,0.15828088919321695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,65536,0.7329004605611166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,8192,0.03398933344417148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,64,0.011443555355072021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,256,0.004547555413511065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,1024,0.028560890091790095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,5120,0.013682666752073499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,8192,0.1633733378516303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1024,10240,0.018087999688254464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,51200,1.2009315490722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,51200,0.2165351046456231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,128,0.006976889239417181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,2560,0.02498133314980401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,5120,0.1299146678712633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,8192,0.07988622453477648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,2560,0.012331555287043253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,8192,0.09205955266952515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,128,0.03930844532118903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,2560,0.014087999860445658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,4096,0.6132328775193956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,768,0.011666666302416058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,128,0.0027875554644399216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,128,0.022869333624839783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,32,0.00997155573632982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,3584,0.01608888970481025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,5120,0.022878222995334204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,512,0.03145600027508206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,6144,0.0058462222417195635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,7168,0.21843022770351836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,65536,0.5602142016092936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,512,0.03814222084151374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,128,0.009444444543785518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,768,0.012720888687504662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,1536,0.030957334571414526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,4096,0.01457244488928053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,2560,0.011425777441925473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,64,0.00813155538505978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,2048,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,16384,0.1766648954815335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,3584,0.44574398464626735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,5120,0.009494221872753566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,1024,0.0353999998834398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,32,0.015080889066060385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,8192,0.12688088417053223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,6144,0.010133333504199982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,6144,0.30817421277364093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,12288,0.04468977782461378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,2560,0.01125955581665039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,32,0.009108444054921469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,1536,0.006520888871616787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,1536,0.12262933784061009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,3072,0.8463422457377116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,32,0.007864889171388414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,65536,5.805641598171658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,256,0.09802666637632583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,6144,0.01332622187005149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,7168,0.04257066713439094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,5120,0.078311112191942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,65536,0.12321333090464275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,256,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,7168,0.3267048994700114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,128,0.005479111025730769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,10240,3.082669364081489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,6144,0.30082665549384224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,8192,1.5592222213745117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,3072,0.024604444702466328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,1536,0.030750221676296655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,1024,0.03283111254374186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,5120,0.011585777832402123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,2560,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,16384,0.21063821845584443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,12288,0.06676355335447523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,3584,0.019508444600635104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,5120,0.02237244447072347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,128,0.0032497776879204642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,7168,0.12681333223978677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,51200,0.06236266427569919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,10240,5.716134389241536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,12288,0.1827982266743978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,1024,0.003976888954639435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,128,0.004932444542646408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,12288,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,12288,0.0765742195977105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,8192,0.02218399941921234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,65536,0.12494222323099773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,7168,0.15067466100056967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,10240,0.9801858266194662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,32,0.02607288956642151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,3584,0.5109715461730957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,12288,0.030078222354253132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,51200,0.06659999820921156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,1024,0.007174222005738153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,4096,0.009190222455395592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,64,0.002776888923512565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,4096,0.006319111006127463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,8192,0.21619821919335258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,6144,0.07493155532413058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,1024,0.005423111220200856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,128,0.007567110988828871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,16384,0.11212000581953262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,256,0.19864443937937418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,256,0.003154666680428717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,5120,0.011576889289749993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,64,0.004366222355100843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,2048,0.039763556586371526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,128,0.00553688903649648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,8192,0.021701332595613267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,6144,0.02682488825586107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,65536,0.06255910793940227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,4096,0.013062222136391534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,65536,0.015234667393896313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,7168,0.032961779170566134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,10240,0.026910222238964502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,1024,0.005968889014588461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,6144,0.07087022066116333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,3072,0.036128888527552284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,65536,0.07607822285758124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,6144,0.0105253333846728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,12288,0.027448000179396734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,32,0.005728000154097875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,768,0.012562666502263812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,1536,0.08225155538982816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,256,0.0036799998746977914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,512,0.006056889063782162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,7168,0.03105688757366604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,2048,1.2896639506022136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,3584,0.012096888489193387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,5120,0.07158578104443021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,1536,0.2454462316301134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,128,0.00699733363257514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,1536,0.005613333235184352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,8192,0.03298755486806234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,3584,0.010293333066834344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,64,0.0033093332830402586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,65536,0.13996888531578913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,128,0.004416888786686791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,64,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,2048,0.031055112679799397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,2048,0.008566222257084316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,10240,0.05517066518465678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,4096,0.019740444090631273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,3072,0.008114666574531132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,2048,0.00850577818022834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,2048,0.0539173318280114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,32,0.03674488928582933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,1024,0.02125155594613817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,32,0.019711999429596793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,32,0.002624888800912433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,3072,0.6618853145175511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,64,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,4096,0.014021333720948962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,3072,2.9088870154486766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,512,0.03376177615589566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,16384,0.06530221965577868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,10240,0.08157333400514391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,5120,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,4096,0.2109768920474582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,768,0.01814933286772834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,2560,0.013660444153679742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,512,0.007276444799370236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,5120,0.07972888814078437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,16384,0.1518204477098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,16384,0.23590487904018828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,4096,0.01682044400109185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,3584,1.0383324093288846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,65536,6.1912935045030375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,3584,0.15796977943844265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,4096,0.01997422178586324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,768,0.037978665696250066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,3072,0.01737422247727712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,1024,0.005158222383923001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,768,0.011644444531864591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,12288,0.027229333917299908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,16384,0.4443511168162028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,51200,4.35135989718967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,32,0.016629333297411602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,3072,0.021928000781271193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,7168,0.03430044319894578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,6144,0.02493688960870107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,2048,0.4239822228749593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,2048,0.017843556072976854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,12288,0.027757333384619817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,512,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,3072,0.014288889037238227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,12288,3.6129468282063804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,3072,0.2480391131507026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,1024,0.006809777683681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,2048,0.05959733327229818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,2560,0.008044444852405125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,32,0.007291555404663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,32,0.12232532766130234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,512,0.004913777940803104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,2560,0.010501333408885531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,2560,0.006335111127959357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,64,0.21514311101701525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,6144,0.1980959971745809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,51200,1.1069884830050998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,10240,0.01664622293578254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,128,0.00589155571328269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,2560,0.011712888876597086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,12288,0.12936977545420328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,1536,0.012551999754375882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,768,0.005943999936183293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,65536,0.0692382189962599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,1536,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,256,0.013292444248994192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,16384,0.40229598681132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,10240,0.34296711285909015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,5120,0.13800177309248182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,64,0.005616000129116907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,8192,0.06007555458280775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,1536,0.2147324482599894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,2560,0.02755644420782725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,4096,0.05876266956329346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,51200,0.054261333412594266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,12288,0.01316088851955202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,10240,0.026696888936890498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,10240,0.10702044434017605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,10240,0.018292443619834054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,7168,0.8337155448065864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,16384,0.4722115728590224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,2048,0.0064080001579390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,1536,0.018657777044508193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,64,0.011213333242469363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,7168,0.014814222852389017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,256,0.004148444367779626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,16384,0.006606222026877933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,4096,0.014174222946166992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,64,0.006785777707894643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,8192,0.16007467110951742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,5120,0.014347554908858405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,256,0.004960888789759742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,32,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,1536,0.011016888750924004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,64,0.008227555288208855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,512,0.003702222059170405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,128,0.013779555757840475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,768,0.008209777375062307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,768,0.016952888833151925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,1024,0.006719999843173557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,3584,1.2913120057847765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,16384,0.04059200154410468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,51200,1.3386418024698894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,768,0.0236151119073232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,3584,0.035170667701297335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,7168,0.05921600262324015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,3072,0.9640862147013346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,2560,0.08879466851552327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,51200,0.3660506672329373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,1536,0.08749155865775214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,32,0.003440888805521859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,3584,0.02234399980968899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,51200,0.4861609141031901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,65536,0.1809742185804579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,32,0.005038222091065513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,65536,0.22306489944458008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,1024,0.00960000024901496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,10240,2.1885484059651694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,1536,0.007102222078376346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,256,0.0069324444565508105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,65536,0.12201600604587132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,16384,0.07538400093714397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,12288,0.006679111056857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,8192,0.009298666483826106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,256,0.018906666172875296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,65536,0.5534124374389648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,3072,0.08496622244517009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,10240,0.18025867144266763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,12288,0.053640888796912305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,7168,0.08216710885365804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,2560,0.012888000243239932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,4096,2.1484781901041665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,16384,0.3132195472717285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,10240,0.1421484417385525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,1536,0.009957333405812582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,6144,0.07905155420303345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,6144,0.2638853391011556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,2048,0.022170666191313002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,32,0.0030942221896515954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,65536,0.8941556082831489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,512,0.006986666884687211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,7168,0.0473671125041114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,7168,0.014518222875065274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,8192,0.015075556106037565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,3072,0.027160889572567407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,51200,0.06536355283525255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,1536,0.00795377790927887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,5120,0.7180950906541613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,2560,0.09705511066648696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,8192,0.9937991036309136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,512,0.005712000032265981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,1024,0.007176888485749562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,768,0.039870222409566246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,768,0.006188444379303191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,12288,0.028798222541809082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,16384,0.022520000735918682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,6144,0.01054133309258355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,2560,0.035001777940326266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,12288,0.006578666468461354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,65536,0.8971591525607638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,3072,0.25817423396640354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,128,0.006809777683681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,4096,0.028049776951471966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,512,0.12170399559868707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,10240,0.015986666083335876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,10240,0.01738844480779436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,3072,0.01329955541425281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,768,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,128,0.012424888710180918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,256,0.0063164445261160536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,2560,0.017841777867741056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,512,0.010536889235178629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,3584,0.028685334655973647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,2560,0.021889777647124395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,512,0.005320888840489917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,8192,0.015075556106037565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,128,0.00620888877246115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,32,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,1536,0.4567759831746419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,5120,0.38205954763624406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,256,0.003160888950030009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,32,0.5207804573906792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,10240,0.021560000048743352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,1536,0.0449848903550042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,51200,0.10219022300508286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,10240,0.019152889649073284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,3584,0.021368000242445204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,128,0.005097777893145879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,3072,0.10840355687671238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,16384,0.02343911098109351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,8192,0.00872444444232517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,7168,0.013951111170980664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,128,0.0069448889957533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,2560,0.06558222240871854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,16384,0.0941351122326321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,2560,0.007140444384680853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,1536,0.012111110819710625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,4096,0.016191111670600045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,64,0.008519111408127679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,12288,0.23576445049709746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,3072,0.010252444280518426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,2560,0.012167111039161682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,6144,0.03140177660518222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,65536,0.1723724471198188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,4096,0.009031111167536842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,1536,0.0058959999846087555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,6144,0.07730222410625882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,512,0.01629955569903056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,1024,0.06604355573654175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,512,0.00517155561182234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,1024,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,2048,0.014575999644067554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,12288,1.860513687133789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,2560,0.02606133288807339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,8192,1.424666616651747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,6144,0.30514666769239635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,256,0.010075555907355415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,3584,0.010303110712104374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,4096,0.0780151089032491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,128,0.005892444401979446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,1024,0.0041271112859249115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,256,0.00432977784011099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,512,0.0103511114915212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,1024,0.02552088929547204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,65536,0.39258845647176105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,1536,0.005969777703285217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,5120,0.16808533668518066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,6144,0.20979377958509657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,2048,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,512,0.005674666828579373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,6144,0.01663555536005232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,12288,0.02094399929046631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,12288,0.08044266700744629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,1536,0.01999288962946998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,6144,0.026171555121739704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,2048,0.0572320024172465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,10240,5.10440190633138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,6144,0.18701956007215711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,64,0.011050666371981302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,512,0.011614222493436603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,16384,0.05951911211013794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,256,0.0041857775714662345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,3584,0.27131199836730957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,5120,0.06631377670500013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,4096,0.054478221469455294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,2048,0.013441777891582914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,64,0.013016888664828407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,16384,0.08532444636027019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,4096,0.257496886783176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,2560,0.010054222411579555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,51200,6.590236239963108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,6144,0.1084444456630283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,16384,0.034197333786222674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,12288,0.0207368897067176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,5120,0.04387733340263367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,12288,0.05196711089875963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,3584,0.09159911341137356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,1536,0.04668088754018148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,16384,0.045737776491377086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,512,0.007064888874689738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,5120,0.02349600030316247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,8192,5.6121643914116754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,12288,5120,0.26980090141296387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,3072,0.009945777555306753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,4096,0.03658488723966811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,12288,0.0796604421403673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,512,0.006083555519580841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,4096,0.5616728994581435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,1536,0.08637244171566433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,6144,0.2670026620229085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,5120,0.007088888850477006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,7168,0.009708444277445475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,3584,0.007893333004580604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,128,0.0032844444115956626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,8192,0.026771555344263714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,10240,0.03718044360478719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,2048,0.0069155556460221606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,16384,0.2778888808356391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,2560,0.012620444099108378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,2560,0.007653333246707916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,2560,0.008273777862389883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,6144,0.019861333900027804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,32,0.051593777206208974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,6144,0.11260088284810384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,12288,0.05996355745527479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,8192,0.020911999874644812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,4096,0.056919111145867236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,1536,0.11007110940085517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,3584,0.015584889385435315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,7168,0.04555733336342705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,65536,0.4989297654893663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,65536,0.03243555625279745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,8192,0.052928000688552856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,5120,0.018522666560279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,256,0.009737778041097853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,10240,0.15597689151763916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,8192,0.05365955498483446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,768,0.003629333443111844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,16384,0.3111084567175971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,1536,0.031016889545652602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,1024,0.00608711110221015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,51200,0.2293173207177056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,768,0.018458666072951425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,768,0.02349777850839827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,12288,2.609500461154514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,8192,0.22739201121860078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,10240,0.10385421911875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,7168,0.1395217842525906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,32,0.007458666960398356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,51200,0.13425689273410374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,768,0.043904887305365674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,2048,0.013901333014170328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,10240,0.021029333273569744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,1536,0.01311644415060679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,512,0.020788444413079154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,2560,0.01344622257683012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,4096,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,32,0.01627288924323188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,65536,0.12655733691321477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,6144,0.04130666785769992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,256,0.006595555692911148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,2048,0.007909333540333642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,7168,0.011153777440388998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,3584,0.1180373297797309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,1024,0.03720000055101182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,8192,0.02293066680431366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,2048,0.04665689004792107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,32,0.02572088936964671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,256,0.006301333506902059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,6144,0.20546933015187582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,7168,0.031103110975689355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,3072,0.011387555963463254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,1024,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,1536,0.05220266514354282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,768,0.0036355555057525635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,3072,0.02429955535464817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,8192,3.5535066392686634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,7168,1.4150364134046767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,5120,0.03196622265709771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,7168,0.19588889016045463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,2560,0.013208000196350945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,2048,0.00923288861910502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,51200,0.07509511046939425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,51200,0.2683582305908203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,1536,0.0662613312403361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,16384,0.5773528946770562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,128,0.010703110860453712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,6144,0.0070053330726093715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,256,0.003656889001528422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,7168,0.529954645368788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,2560,0.028841776980294123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,64,0.0029884444342719186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,768,0.004038222134113312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,64,0.002642666714058982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,8192,0.6504711045159234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,8192,0.025812443759706285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,5120,0.06286222404903836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,256,0.006672888994216919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,51200,0.10997422536214192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,8192,0.05650044149822659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,256,0.01657777859105004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,8192,0.09206933445400661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,4096,0.008169777691364288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,65536,0.5854310989379883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,3072,0.20494488875071207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,12288,0.03868622250027127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,8192,0.006963555183675554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,768,0.016647110382715862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,64,0.0032186667538351486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,128,0.028395556741290625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,10240,0.04957155717743767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,64,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,3072,0.006211555666393704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,256,0.0032515554792351196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,5120,0.10274489058388604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,3584,0.006341333190600078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,7168,0.18011555406782362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,2048,0.025219556358125474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,3584,0.20745600594414604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,5120,0.29085689120822483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,1024,0.03948355714480082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,3584,0.009152000149091085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,51200,0.06951555278566149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,256,0.005709333138333426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,5120,0.007902222375075022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,32,0.008641777767075432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,16384,0.07598844501707289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,64,0.0030782222747802734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,65536,0.08143822352091472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,7168,1.1128898196750217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,1024,0.008229333493444655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,5120,0.07070222165849474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,5120,0.014176888598336114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,1024,0.03167377909024557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,6144,2.8004258473714194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,64,0.017017778423097398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,128,0.012859555582205454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,3072,0.021366222037209406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,51200,0.20497688982221815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,1024,0.007877333296669854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,4096,0.38250666194491917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,8192,256,0.0074675555030504865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,51200,0.3802631166246202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,65536,1.8228160010443792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,65536,0.37016889784071183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,12288,0.03755733370780945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,512,0.008513777620262569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,3072,0.010885333021481832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,128,0.00665244460105896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,1536,0.009457777771684859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,3584,0.011395555403497485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,65536,0.42528888914320206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,3072,0.014141332772043018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,7168,0.1600195566813151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,10240,0.07960622178183661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,1536,0.0072266666425599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,1536,0.07020888725916545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,2048,0.00739644467830658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,51200,0.1575288905037774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,51200,0.05341866612434387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,512,0.03815911213556925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,512,0.00811644477976693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,256,0.18698044617970785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,10240,0.10223466820187038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,5120,0.03151822090148926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,7168,0.05442400111092461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,128,0.009175111022260454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,1024,0.1109422246615092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,6144,0.02780533168050978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,512,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,2048,0.006721777634488211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,7168,0.02838133441077338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,5120,0.055517332421408765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,8192,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,7168,0.7027991082933215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,3584,0.10636533631218804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,3584,0.2952977816263835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,51200,0.07158488697475858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,51200,0.6123929023742676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,51200,0.014289778139856128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,3072,0.07822844717237684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,768,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,6144,0.11991733974880642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,12288,0.05570399761199951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,32,0.0026168889469570587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,8192,0.27470400598314076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,12288,0.06771466467115614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,512,0.007915555602974361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,64,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,512,0.004423111263248655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,4096,0.007608888877762689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,3072,0.020593777298927307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,768,0.005406222409672207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,1024,0.01519555515713162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2048,32,0.005323555734422471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,6144,0.02569688856601715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,8192,0.5365911059909397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,768,0.004592888885074192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,1024,0.03965866565704346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,768,0.01622577839427524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,3072,0.3421946631537543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,7168,0.015040000279744467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,10240,0.1550017727745904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,65536,0.0781680014398363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,12288,0.02616888946957058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,3584,0.03203644355138143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,128,0.002921777880854077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,256,0.010684444672531553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,1024,0.07268799675835504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,3584,0.007334222396214803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,8192,1.2329235076904297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,16384,0.9663591384887695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,16384,0.2723982334136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,51200,0.04356355468432108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,2560,0.0069022224181228215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,32,0.0029075555503368378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,3072,0.01439377831088172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,128,0.00544888898730278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,51200,0.17079555988311768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,65536,0.3746346632639567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,10240,0.17393867174784342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,51200,0.07576088772879706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,1024,0.0047688889834615915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,6144,0.5372311274210612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,3584,0.08640799919764201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,5120,0.2682480017344157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,16384,0.514119995964898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,7168,0.14517866240607366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,7168,1.7371848424275715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,768,0.052054223087098866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,5120,0.050985776715808444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,10240,0.24390665690104166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,12288,0.016107555892732408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,51200,0.06893689102596708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,2048,0.01254488858911726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,51200,0.3701919979519314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,12288,0.08986310826407538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,32,0.0029004443850782183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,65536,0.07534933090209961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,768,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,1024,0.0058240000572469496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,16384,0.007017778025733099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,3584,0.05351288782225715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,6144,0.02548177705870734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,2560,0.007115555306275685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,1024,0.007524444825119442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,1536,0.08086755540635851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,5120,0.007060444189442529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,7168,0.03640888796912299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,7168,0.02603200078010559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,6144,0.09628888633516099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,3584,0.00812533332241906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,2560,0.01901155544651879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,3584,0.08572444650861953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,256,0.007939555578761632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,4096,0.012340444657537671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,6144,0.013561777770519257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,16384,0.038959999879201256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,128,0.029531555043326482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,8192,0.020695999264717102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,128,0.00300799993177255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,16384,0.20332354969448516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,12288,0.20636710855695936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,16384,1.1089626948038738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,3584,0.12036889129214817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,3072,0.009233777721722921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,5120,0.016741333736313712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,16384,0.07960800329844157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,2560,0.056996444861094155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,128,0.004046222401989831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,512,0.003690666622585721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,3584,0.13533066378699407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,768,0.3755164411332872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,4096,0.017795556121402316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,2560,0.00723555518521203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,5120,0.19890666007995605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,512,0.012738666600651212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,3072,0.3978488975101047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,7168,0.04344800114631653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,7168,0.013907555076811047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,4096,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,256,0.00812000036239624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,512,0.06723911232418485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,8192,0.32005156411064994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,10240,0.0138586668504609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,2560,0.006627555522653792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,32,0.005183111048407025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,512,0.014334221680959066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,65536,0.18666488594479033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,16384,0.07798044549094306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,16384,0.039376888010236956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,12288,0.6861279805501302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,1536,0.007466666400432587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,768,0.030769778622521296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,4096,0.010973333484596677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,128,0.0035946667194366455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,2048,0.011726222104496427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,128,0.004835555536879433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,3072,1.1048399607340496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,12288,0.008040000167157915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,3584,0.03740088807211982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,6144,17.11437818739149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,2048,0.08128977484173246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,5120,0.00979644474056032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,6144,0.015632000234391954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,512,0.006615110983451207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,1024,0.04658755660057068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,2560,0.011554666691356234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,64,0.005095110999213325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,8192,0.03298399845759074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,10240,0.04028266668319702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,65536,0.9182444678412544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,4096,0.0105262224872907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,32,0.02522755497031742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,256,0.002924444360865487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,7168,0.018403554956118267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,1024,0.006480000085300869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,6144,3.1269412570529513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,2048,0.09487555424372356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,512,0.0321342216597663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,256,0.012640000217490725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,3072,0.05428444345792135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,512,0.006124444305896759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,10240,0.02942133280966017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,2560,0.13681422339545354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,5120,0.2534231079949273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,4096,0.07366933425267537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,16384,0.06396444638570149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,65536,1.0432728661431205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,2048,0.06724888748592801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,6144,0.026156443688604567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,3584,0.017834666702482436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,12288,0.036532445086373225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,6144,0.09461600250667995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,6144,0.027335999740494624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,128,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,51200,0.025415110919210646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,10240,0.06147378020816379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,16384,0.012132444315486483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,256,0.004293333325121137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,8192,0.17418932914733887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,6144,0.01019377758105596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,10240,0.3570515579647488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,5120,0.007618666523032718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,256,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,8192,0.051608890295028687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,65536,0.1759288840823703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,2048,0.009564444422721863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,16384,4.789368947347005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,3584,0.053555554813808866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,256,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,128,0.03304800060060289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,2048,0.015444444285498725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,6144,0.10793333583407932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,512,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,256,0.005931555396980709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,768,0.006712889091836081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,7168,0.15167110496097141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,3584,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,10240,0.11493333180745442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,4096,0.010298666854699453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,1024,0.01127822200457255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,6144,0.011857777833938599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,2560,0.012716444830099741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,8192,0.08512266476949056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,51200,0.09027110868030125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,8192,0.04045777850680881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,51200,0.05010133319430881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,2048,1.4140621821085613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,1536,0.013153778182135688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,65536,0.08884888887405396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,256,0.003043555551105075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,12288,0.2371137672000461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,51200,0.42953154775831437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,4096,0.011010666688283285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,16384,0.5227155685424805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,1024,0.011599111060301462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,1024,0.004320888883537716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,5120,0.8831413586934408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,3584,0.013207111093733044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,768,0.007121777368916406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,64,0.004004444513056013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,8192,0.015511110424995422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,3072,0.07768266730838351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,768,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,8192,0.19704088899824354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,128,0.009528888596428765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,512,0.012856889102194043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,7168,0.00849333322710461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,12288,0.016395555602179635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,1536,0.05919733312394884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,3584,0.012621333201726278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,5120,0.04647644360860189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,128,0.011789333489206104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,256,0.028925332758161757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,256,0.014667555689811707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,768,0.016336000627941556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,51200,0.16090933481852213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,16384,0.02534666657447815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,3584,0.006909333169460297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,2048,0.08892977899975246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,1536,0.010513777534166971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,8192,0.03363555669784546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,10240,0.01111288865407308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,51200,0.40531465742323136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,8192,0.027439999911520217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,256,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,2048,0.035303112533357404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,65536,0.10663111342324151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,4096,0.009315555294354757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,5120,0.42053688897026914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,65536,0.23223288853963217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,512,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,8192,0.044383111927244395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,5120,0.023736000061035156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,7168,0.020198222663667467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,65536,1.4178844028049047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,512,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,512,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,128,0.005187555733654234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,4096,0.015757333901193406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,2560,0.1568160057067871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,5120,0.03602755400869582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,768,0.00666222224632899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,10240,0.00870577742656072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,128,3072,0.0069822221994400024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,256,0.003456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,64,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,51200,0.15221244759029812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,512,0.004970666848950916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,65536,0.365064885881212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,2560,0.008196444147162968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,12288,0.22601866722106934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,16384,0.03159644537501865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,128,0.00564177789621883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,12288,0.02128711177243127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,1024,0.10204799969991048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,2560,0.2561911212073432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,1536,0.08245599932140775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,7168,0.025803556044896443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,1536,0.010044444766309526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,128,0.0028293333533737394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1536,3584,0.023364444573720295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,10240,0.14480533864763048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,128,0.0027724444452259275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,128,0.06920799944135878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,1024,0.005550222264395819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,128,0.003626666549179289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,12288,0.02275288932853275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,2048,0.013557333085272046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,512,0.006028444402747684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,16384,0.13856889141930476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,256,0.005479111025730769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,1024,0.35155555937025285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,65536,0.0707084470325046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,12288,0.3157484531402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,65536,0.06052177482181125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,1024,0.04938844508594937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,768,0.07172799772686429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,3584,0.008112000094519721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,6144,2.7096631791856556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,51200,0.7673741976420084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,512,0.007872888611422645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,7168,0.2890995608435737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,2560,0.051426665650473706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,8192,0.013616888887352414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,3072,0.009782222410043081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,3584,0.008045333127180735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,12288,0.498576905992296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,3072,0.00962577760219574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,64,0.010423999693658618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,12288,0.03750133183267381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,256,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,1536,0.0454293323887719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,3584,0.027896000279320612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,16384,0.1454604466756185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,7168,3.9604843987358938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,256,0.007314666277832455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,32,0.003621333175235324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,512,0.018138666947682697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,3584,0.012343111137549082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,768,0.021168000168270532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,51200,5.4381459554036455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,7168,0.13990667131212023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,512,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,2560,0.016591111818949383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,1024,0.005704000178310607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,1024,5.595875634087457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,65536,37.10638766818577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,2560,0.7444186740451388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,512,0.003858666867017746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,6144,0.1367644468943278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,1024,0.006424888968467712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,4096,0.1083822250366211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,6144,0.020216888851589628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,768,0.010220444036854638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,128,0.006851555572615729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,3072,0.0417520006497701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,128,0.007263110743628607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,51200,0.013986666997273764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,6144,0.01070666644308302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,65536,0.38100534015231663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,2560,0.0609804458088345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,3584,0.02532177832391527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,51200,0.12745066483815512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,64,0.3299502266777886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,32,0.011787555283970304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,51200,0.060917331112755664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,6144,0.019374221563339233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,768,0.01960000064637926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,12288,0.05219911204444038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,4096,0.10888533459769355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,7168,0.02472355630662706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,768,0.03551466597451104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,3072,0.018837332725524902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,128,0.16328267256418863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,4096,0.32025334570142955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,2048,0.119504001405504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,7168,0.016945777667893302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,7168,0.454570664299859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,10240,0.19171999560462105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,1536,0.005012444323963589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,4096,0.09981511036554973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,32,0.006871999965773688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,12288,0.0461751123269399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,51200,0.08672800328996445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,65536,0.8005075454711914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,65536,2.767731560601128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,512,0.019712888532214694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,10240,0.05584088961283366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,2560,0.015020444989204407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,1536,0.007312888900438945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,7168,0.01963644391960568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,5120,2.9816409216986766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,6144,1.147125350104438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,12288,0.5405111312866211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,8192,0.02149155570401086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,256,0.016177778442700703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,1536,0.12024977472093369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,51200,3.055437299940321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,65536,0.19122666782803008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,32,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,16384,0.03801511062516107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,4096,0.015253333581818474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,6144,0.02732266651259528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,10240,0.06666400035222371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,1536,0.028154667880800035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,1024,0.004597333156400257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,8192,0.32711021105448407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,65536,0.1714320050345527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,32,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,7168,0.014239999983045789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,1536,0.03284800052642822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,51200,0.2337484359741211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,4096,0.01035288886891471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,3584,0.6739724477132162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,768,0.018579555882347953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,2048,0.06568355692757501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,16384,0.04001155495643616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,8192,0.383025778664483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,65536,0.620566209157308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,256,0.032356444332334734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,3584,0.08064089218775432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,1536,0.006162666612201267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,3584,0.05724177757898966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,768,0.08026844263076782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,7168,0.12328710820939805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,3072,0.204130662812127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,6144,0.017050666941536795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,5120,0.013364444176355997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,768,0.0053680001033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,8192,0.027650665905740526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,1024,0.023504000571038988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,1024,0.047798222965664335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,65536,0.9204186333550347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,4096,0.04384266667895847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,6144,0.01589777734544542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,128,0.01779111060831282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,256,0.005029333134492238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,2048,0.073125335905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,65536,0.2382915549808078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,1024,0.005728888842794631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,6144,0.01371911085314221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,7168,0.0233297778500451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,64,0.004678222040335338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,12288,0.04289244280921089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,10240,0.18886044290330675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,768,0.006018666757477655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,1024,0.0058373332851462895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,128,0.006474666711356904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,2048,0.013432888521088494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,32,0.0033528889632887314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,12288,0.009248889154858058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,3072,0.007522666619883643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,1024,0.0041751112374994485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,51200,0.26452088356018066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,64,0.15276088979509142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,7168,1.6922444237603083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,2048,0.023342221975326538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,5120,0.08760888708962335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,256,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,3072,0.01759288873937395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,10240,0.04712355468008253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,1536,0.004868444469239977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,128,0.004646222210592694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,5120,0.009404444032245213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,51200,15.912789238823784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,12288,0.01938222183121575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,3584,0.08090844419267443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,3072,0.026923555466863845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,4096,0.02753688891728719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,51200,0.34790934456719297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,3072,0.007726222276687622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,65536,0.07717244492636786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,128,0.00272444449365139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,2560,0.012331555287043253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,2048,0.009884444375832876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,64,0.003485333174467087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,5120,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,512,0.005063999858167437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,5120,0.020302222834693063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,16384,0.15493067105611166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,8192,0.04696711235576206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,10240,0.1409511168797811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,4096,0.011019555230935415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,51200,0.05284711056285434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,256,0.01292622254954444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,6144,0.05232977867126465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,5120,0.07423910829755995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,10240,0.06499733527501424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,256,0.1920551061630249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,2048,0.28044178750779897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,256,0.005254222287072076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,512,0.017070222232076857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,512,0.007455999652544658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,6144,0.03072355522049798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,512,0.03534311056137085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,6144,3.439905802408854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,3584,0.17226933108435738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,768,0.004012444367011388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,128,0.016149333781666227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,512,0.042019555966059365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,12288,0.11400800281100804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,3072,0.03485866718822055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,64,0.020310221446885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,16384,0.05760711431503296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,65536,6.19286388821072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,8192,0.018868444694413077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,5120,0.017176000608338248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,512,0.05485688977771335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,65536,0.2942933241526286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,8192,0.024546666277779475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,16384,0.01868711080816057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,10240,0.036533332533306546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,51200,0.6311066415574815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,768,0.020678222179412842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,5120,0.05695555607477824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,768,0.022064000368118286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,8192,0.9004044002956815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,128,0.0034293333689371743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,5120,0.015581332974963717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,51200,0.2690977785322401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,16384,0.026155556241671245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,2560,0.05935644441180759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,12288,0.2361911137898763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,6144,0.018345778187115986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,32,0.007895111209816402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,65536,0.07583466503355238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,64,0.005183111048407025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,512,0.3598346710205078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,3584,0.06463644239637586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,65536,0.08502399921417236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,8192,0.010095111197895473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,1024,0.018936889039145578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,16384,0.03981688949796889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,2560,0.03898310992452834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,65536,1.2453022003173828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,3584,0.08135288953781128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,16384,0.7948275672064887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,5120,0.01201777739657296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,32,0.0037048889531029593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,8192,0.21956888834635416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,12288,0.08371733294592963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,65536,0.23542488945855033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,16384,0.3175537851121691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,1536,0.0428693327638838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,12288,0.1948417822519938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,3584,0.018301332990328472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,1024,0.014327999618318347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,1024,0.010834666589895884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,2048,0.13876444763607448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,2560,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,3584,0.022825777530670166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,512,0.028283556302388508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,32,0.014936000108718872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,256,0.2572702301873101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,6144,0.013943110903104147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,3072,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,6144,0.01831644442346361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,7168,0.05242844422658285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,1024,0.007923555870850882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,32,0.002924444360865487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,6144,0.011318222516112857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,32,0.005260444349712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,1024,0.05535022086567349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,128,0.11027466588550144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,256,0.003036444385846456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,12288,0.1129759947458903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,3584,0.0883457793129815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,8192,0.14588356018066406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,12288,0.02232711182700263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,512,0.03390399946106805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,256,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,1536,0.23322044478522408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,65536,6.812102423773871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,3584,0.04924799998601278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,32,0.005026666654480828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,5120,0.033174223370022245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,128,0.008352888955010308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,768,0.01586577792962392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,768,0.010678222609890832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,2048,0.00776711106300354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,12288,0.034585777256223894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,4096,1.1543199751112196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,768,0.6905493206448026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,128,0.008100444244013892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,51200,0.035000890493392944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,256,0.004625777817434735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,768,0.004001777619123459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,32,0.0058044443527857465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,512,0.0076044441925154785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,2048,0.041414221127827965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,768,2048,0.07352444198396471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,12288,0.031344887283113264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,16384,3.2961343129475913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,65536,0.08298577864964803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,2048,0.008376888930797577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,8192,0.20801333586374918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,2560,0.018315555320845712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,6144,0.0110871113008923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,7168,0.058598220348358154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,16384,0.08996888664033677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,128,0.0041706665522522396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,2560,0.007988444632954067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,1536,0.3020728958977593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,4096,0.014391111003028022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,51200,0.8937706417507596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,3072,0.024697777297761705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,8192,0.007207999626795451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,12288,0.046894222497940063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,2560,0.008280889027648503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,64,0.015785778562227886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,3072,0.007296889192528195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,8192,0.10835288630591498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,2048,0.009487999810112847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,512,0.018012444178263348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,6144,0.020364445116784837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,3072,0.010975999964608086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,51200,0.3752435578240289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,1024,0.004111111164093018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,3072,0.009528888596428765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,3584,0.020410666863123577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1024,51200,0.40499377250671387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,3072,0.009874666730562845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,4096,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,12288,0.050195554892222084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,10240,0.0689493351512485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,6144,0.4018346733517117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,5120,0.07342222001817492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,256,0.005290666802061929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,7168,0.018985778093338013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,12288,0.09043466382556492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,7168,0.0336168905099233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,4096,0.28935199313693577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,128,0.022907555103302002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,65536,0.5309039751688639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,7168,0.021133333444595337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,12288,0.040532443258497454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,512,0.024784889486100938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,1024,0.010919999745157031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,8192,0.035290665096706815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,51200,0.5565368864271376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,2560,0.007982222570313348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,51200,0.5267484452989366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,4096,0.029442667961120605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,1024,0.005177777674463061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,768,0.018966222802797954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,12288,0.14030310842725965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,3072,0.030759112702475652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,1024,0.006712889091836081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,2560,0.016223111086421542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,65536,0.015606222881211175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,3584,0.007989333735571967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,2048,0.07524710893630981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,128,0.0029404444826973807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,4096,0.35019556681315106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,10240,0.007666666474607255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,1536,0.029071112473805744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,3072,0.2761209011077881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,7168,0.016234666109085083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,12288,2.3632435268825955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,6144,6144,0.030582222673628066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,6144,0.0885617799229092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,5120,2.5932178497314453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,16384,12.364611307779947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,10240,0.024877332978778418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,12288,0.1149866713417901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,128,0.008255999949243333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,768,0.011559111376603445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,5120,0.5384053124321831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,65536,0.17443645000457764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,128,0.0028924445311228433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,128,0.002867555452717675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,768,0.0038782221575578055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,65536,4.660767025417752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,1024,0.022692445251676772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,65536,0.4208195474412706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,256,0.027247111002604168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,16384,0.06924266947640313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,10240,2.213647206624349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,16384,0.026893332600593567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,10240,0.019558222757445443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,16384,0.4185573259989421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,3072,0.03541244400872125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,512,0.00816000004609426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,1024,0.004889777965015835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,65536,0.29635734028286403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,10240,0.18310044871436226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,1536,0.010028444230556488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,51200,0.6681866645812988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,10240,0.12054488393995498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,128,0.013360000318951078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,65536,1.1580497953626845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,65536,62.1728515625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,2560,0.29199645254347056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,64,0.00721155520942476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,8192,1.517416000366211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,768,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,1024,0.006722666737106111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,128,0.01867377758026123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,64,0.00299466670387321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,12288,0.019987555013762582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,3584,0.009450666606426239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,5120,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,51200,0.2846248944600423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,1536,0.015056888262430826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,256,0.0033484444850020935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,256,0.006832888970772426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,12288,5.534889645046658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,8192,0.10072800185945298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,768,0.02184444500340356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,512,0.008387555678685507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,8192,0.460918214586046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,5120,0.04177600145339966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,10240,0.14812533060709634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,64,0.005190222213665645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,3072,0.0676835576693217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,65536,0.769992881351047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,8192,0.6250906520419651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,2560,0.008443555898136562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,7168,0.36255378193325466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,51200,0.697898652818468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,7168,0.3247617880503337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,2560,0.007931555310885111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,5120,0.011312888728247749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,1536,0.011980444192886353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,3072,0.06334222025341459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,51200,0.8365715344746908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,51200,0.538348462846544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,2048,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,2560,0.010619555910428366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,6144,0.011342222491900126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,2560,0.11613066991170247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,12288,0.14803377787272134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,32,0.018466666340827942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,256,0.00564533347884814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,2048,0.006640888750553131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,51200,0.5759199990166558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,512,0.03191288974550035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,1024,0.006487111250559489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,256,0.008568888737095727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,8192,0.11076888773176406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,4096,0.12097688515981038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,1024,0.022648889157507155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,2048,0.02104533380932278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,65536,0.06887911425696479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,2560,0.04572711057133145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,6144,0.21602488888634575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,768,0.0038115556041399636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,3584,0.15440977944268122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,4096,0.010325333310498131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,1536,0.06803199979994032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,256,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,768,0.039529777235454984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,65536,0.4779973559909397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,51200,0.0535697771443261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,4096,0.02409244411521488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,128,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,5120,0.01719111038578881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,2048,0.3614319960276286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,768,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,5120,1.1211990780300565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,1536,0.03264000018437704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,4096,0.0634488860766093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,256,0.010040000081062317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,3072,0.11552800072564019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,768,0.02516355613867442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,5120,0.01906755566596985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,7168,2.3131181928846574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,4096,0.014308444327778287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,4096,0.014936889211336771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,128,0.14039644930097792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,16384,0.7715617815653483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,2560,0.03849422269397312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,4096,0.03690133492151896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,256,0.002924444360865487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,3072,0.007633777956167857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,3072,0.14342933230929905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,7168,0.7714755800035266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,1024,0.05705066521962484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,8192,0.11224977175394694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,6144,0.2583128876156277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,3584,0.10674133565690781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,4096,0.030650665362675984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,6144,0.02350577712059021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,3584,0.019133332702848647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,7168,0.04146577914555868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,1536,0.006481777876615524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,256,0.011927111281288994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,1024,0.011487111449241638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,256,0.019205333458052743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,10240,0.07192799780103895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,12288,0.014999111493428549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,256,0.008391111261314815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,10240,0.06798133585188124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,3072,0.05401688814163208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,2560,0.022566222482257422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,1536,0.014236445228258768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,32,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,8192,0.25262223349677193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,512,0.2185911072625054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,12288,0.23737155066596138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,3584,0.08122399780485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,2560,0.007725333174069722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,6144,0.198216888639662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,2560,0.09545689158969456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,128,0.0053280000057485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,16384,0.07556000020768908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,768,0.0069413334131240845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,16384,0.0749982198079427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,12288,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,5120,0.012622222304344177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,10240,0.0443102220694224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,512,0.004851555658711327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,2560,0.008463111188676622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,256,0.01830044388771057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,51200,1.1360933515760634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,12288,0.056346668137444385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,512,0.008031999899281396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,768,0.03230844603644477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,2048,0.009398221969604492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,5120,0.014482667048772177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,51200,0.044396443499459155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,1024,0.006816000160243776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,16384,0.08856266736984253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,64,0.013437333206335703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,512,0.008435555630260045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,768,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,3584,0.12861066394382054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,128,0.00647555540005366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,128,0.015221332510312399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,512,0.013393777940008374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,3584,0.07084000110626221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,5120,0.15782222482893202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,128,0.0053119998839166425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,768,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,4096,0.020743111769358318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,7168,0.05211822191874186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,2048,0.1684417724609375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,64,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,512,0.06337866518232557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,2560,0.04938755432764689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,8192,0.24087288644578722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,5120,0.06292000081804064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,6144,0.08059999677870008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,1024,0.02460177739461263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,65536,0.5896284315321181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,12288,0.29055555661519367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,1536,0.014952888091405233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,3584,0.017603556315104168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,3584,0.10075110859341091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,768,0.004885333279768626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,7168,0.022869333624839783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,6144,0.2751235432094998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,7168,0.04289955563015408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,1024,0.0373253325621287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,16384,0.13017600112491184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,6144,0.031167109807332356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,3072,0.16256177425384521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,256,0.005377777748637729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,32,0.0024782222592168385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,3072,0.434296899371677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,64,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,16384,0.10045778089099461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,12288,0.024667556087176006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,7168,0.1802764468722873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,1024,0.008082666330867344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,51200,0.15909422768486872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,768,0.48911110560099286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,2048,0.012198222180207571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,7168,0.02238311039076911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,2048,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,5120,0.08966844611697727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,128,0.01841600073708428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,128,0.0033617777129014335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,3072,0.008233778178691864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,4096,0.03972622089915805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,12288,0.1712426609463162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,12288,0.9056284162733289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,1536,0.01221333361334271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,1536,0.03321955601374308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,3072,2.0083014170328775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,768,0.010795555180973478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,12288,1.0030782487657335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,3584,0.043699555926852755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,4096,0.09735111395517985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,10240,0.2552293406592475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,512,0.019888000355826486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,3584,0.17966932720608184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,3072,0.019103111492262948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,768,0.025455110602908667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,6144,0.2579982280731201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,512,0.004866666677925321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,128,0.010949333508809408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,6144,0.4963155322604709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,128,0.005108444641033809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,65536,0.12455644872453477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,7168,0.010264889233642155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,2048,0.01034400032626258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,8192,0.1311591068903605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,16384,0.2594257725609673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,4096,0.11894310845269097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,51200,4.68228997124566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,4096,1.0165138244628906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,8192,0.13520088460710314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,16384,0.009545777406957414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,10240,0.030972444348865088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,2560,0.01681777834892273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,1536,0.008329778081840938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,16384,0.14991999997033015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,512,0.020641777250501845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,5120,0.04066577884886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,51200,0.6585120095147027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,1536,0.008003555238246918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,51200,0.049810667832692467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,10240,0.11444177892473008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,1536,0.005712000032265981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,51200,0.06886933247248332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,6144,0.020752888586786058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,768,0.017065778374671936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,3072,0.2694968912336561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,3584,0.11916889084710015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,512,0.0034497777620951333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,3072,0.05813866853713989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,8192,4.206217024061415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,4096,0.01406311160988278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,7168,0.03531644410557217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,512,0.003673777812057071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,8192,0.04484355449676514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,7168,0.01479200025399526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,5120,0.1228826708263821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,3584,0.02613688839806451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,8192,0.05443466703097025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,256,0.0036142224238978494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,6144,0.013752000199423896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,3584,0.024290665984153748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,8192,0.014231999715169271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,256,0.029614223374260798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,64,0.002595555658141772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,6144,0.09358844492170547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,768,0.005423111220200856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,3584,0.07863110966152616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,32,0.0129457778400845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,1536,0.0359919998380873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,7168,0.010991111397743225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,768,0.023564444647894964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,128,0.02846577763557434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,1536,0.8425031238132052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,768,0.010216000179449717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,512,0.004683555414279302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,4096,0.018770666586028207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,16384,0.34542044003804523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,51200,0.7089697519938151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,6144,0.04232800006866455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,5120,0.014736000034544202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,6144,0.08099822203318278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,3072,0.011993777420785693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,5120,0.009754666851626502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,5120,0.08133600155512492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,4096,0.00904177791542477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,6144,0.011653333074516721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,256,0.013722666435771518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,768,0.01534222231970893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,5120,0.08220622274610731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,7168,0.03582311007711623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,51200,0.43684710396660703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,2560,0.09280444516075982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,10240,1.039818657769097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,256,0.030877331892649334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,3072,0.02121777832508087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,5120,0.01427111029624939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,16384,0.40015554428100586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,1024,0.006744888921578725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,768,0.003842666745185852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,12288,0.08468088838789199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,7168,0.050472001234690346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,65536,0.7094568676418729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,4096,0.091812445057763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,51200,0.2649573220147027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,3072,0.007461333440409766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,256,0.005159111072619756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,51200,0.05763022104899088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,10240,0.0181688881582684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,128,0.007177777588367462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,7168,0.24245778719584146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,2560,0.01867288847764333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,768,0.006152888966931238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,4096,0.026928888426886663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,16384,2.0264559851752386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,12288,0.08868622117572361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,12288,0.3919324345058865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,128,0.015609777635998197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,12288,0.02912799848450555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,3072,0.014522666732470194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,5120,0.6079529126485189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,2560,0.12490667237175836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,12288,0.033599111768934466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,128,0.011408000356621213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,256,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,10240,0.10362222459581162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,3584,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,256,0.008257778154479133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,7168,0.16198489401075575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,3072,0.013289777768982781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,4096,0.01942399972014957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,6144,0.7266471121046277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,8192,0.05015288790067037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,6144,0.01772622267405192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,256,0.016918222109476726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,5120,0.01147466649611791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,4096,0.010383110907342698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,1024,0.00574933323595259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,6144,0.036892443895339966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,64,0.005290666802061929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,3072,0.015712888704405892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,51200,3.6040941874186196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,2048,0.00960177762640847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,16384,0.32922132809956867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,65536,0.2717137866550022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,6144,0.030395554171668157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,6144,0.023825777901543513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,768,0.026752000053723652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,12288,0.010464888479974536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,768,0.020581333173645865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,12288,0.6289955774943033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,2560,0.3123137685987684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,1536,0.1506835619608561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,32,0.00739822205570009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,6144,0.0517520010471344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,10240,0.6400284237331814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,32,0.0030577776746617425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,4096,0.03817066550254822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,1024,0.05462755428420173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,65536,0.5598070886400011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,51200,1.6705182393391926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,2048,0.01757688820362091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,6144,0.008045333127180735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,5120,0.0311155551009708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,512,0.0069271110826068455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,2560,0.010031110710567897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,2048,0.024162666665183172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,2048,0.011418666276666852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,6144,0.19657954904768202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,65536,0.09468444188435872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,2048,0.010426667001512317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,7168,0.014672888649834527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,65536,0.03352444370587667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,1024,0.006170666466156642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,64,0.004232888834344016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,4096,0.00612088872326745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,3584,0.05964622232649061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,51200,1.649597379896376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,51200,2.2994808620876737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,512,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,768,0.00889866633547677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,256,0.012551111479600271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,7168,0.023762666516833838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,65536,0.0277173337009218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,4096,0.06009599897596571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,65536,1.3897209167480469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,3584,5.929569668240017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,1536,0.01572800013754103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,6144,0.19920089509752062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,256,0.004015999949640698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,12288,0.05154044429461161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,3072,0.3160826630062527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,12288,0.6281786494784886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,7168,0.022322666313913133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,7168,0.016736888223224215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,512,0.059337774912516274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,8192,0.010315555665228102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,51200,0.08319644133249919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,8192,0.04313333498107063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,6144,0.10002133581373428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,2048,0.04347199863857693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,6144,0.0943911141819424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,10240,0.022651554809676275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,7168,0.02051466703414917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,64,0.0031866667171319327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,128,0.1219866673151652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,12288,0.024534222152498033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,7168,0.056484447585211865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,65536,0.07963999774720934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,10240,0.058713780509101025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,3072,0.862243546379937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,2560,0.07455022467507257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,256,0.011372444530328115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,16384,0.02686311139000787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,12288,0.08369688855277167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,1536,0.012085333466529846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,64,0.007052444749408298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,64,0.003112000102798144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,3072,0.012311110893885294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,5120,0.01292622254954444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,8192,0.028713779317008123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,128,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,8192,0.022879110442267522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,16384,0.5263404316372341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,3584,0.019239111079110038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,5120,0.17086133691999647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,5120,0.023638221952650283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,768,0.009580444130632613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,5120,0.1751404470867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,16384,0.06653155883153279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,5120,0.08082489172617595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,2560,0.14940888351864284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,128,0.005507555686765247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,10240,0.03385511040687561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,12288,0.018951111369662814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,32,0.007233777807818518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,256,0.004416000097990036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,2048,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,10240,0.020729778541458976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,10240,0.032646222247017756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,7168,0.025575111309687298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,3584,0.12627911567687988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,65536,0.1252639955944485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,3072,0.010188444621033138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,2560,0.00981866651111179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,512,0.017621333400408428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,6144,0.015427556302812366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,65536,0.10985777775446574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,10240,0.053716444306903414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,3584,0.0909422238667806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,128,0.0028035555862718155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,51200,1.906818601820204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,256,0.20612800121307373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,8192,0.08236355251736112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,32,0.019083556201722886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,6144,0.14270488421122232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,65536,0.4616035355461969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,128,0.030957334571414526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,512,0.005480888817045424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,1024,0.009324444664849175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,10240,4.811812506781684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,7168,0.40472888946533203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,10240,0.11151733663347031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,2048,0.008739555875460306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,2560,0.07893600066502889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,128,0.017398221625222098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,51200,0.14994133843315974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,8192,1.6019218232896593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,6144,0.4732649061414931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,3072,0.3103422323862712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,3584,0.009528000321653154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,1536,0.020655999581019085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,768,0.003969777789380816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,2560,0.022416889667510986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,16384,0.018212444252438016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,1536,0.011077333655622272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,7168,0.030682666434182063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,51200,0.29648621877034503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,1536,0.07403377691904704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,1024,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,128,0.003351111171974076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,512,0.003956444561481476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,2560,0.006643555644485686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,10240,0.10002933608161078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,1536,0.018959111637539335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,2560,0.01620444489849938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,64,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,4096,0.11284889115227593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,6144,0.5431564648946127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,65536,0.9991448720296224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,5120,0.050605333513683744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,1536,0.006005333529578314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,16384,0.047712001535627574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,128,0.01685333251953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,1536,0.004447111239035924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,3584,0.012479110724396177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,2560,0.007958222594526079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,512,0.005434666656785541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,65536,0.28999378946092397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,7168,0.1614284382926093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,7168,0.017648888958825003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,64,0.036419557200537786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,128,0.005388444496525659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,128,0.02072177827358246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,5120,0.024343111448817786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,1536,0.06496089034610324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,4096,0.009496000077989366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,2048,1.2895431518554688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,3072,0.00963466697269016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,51200,0.10247111320495605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,1024,0.010033778018421596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,1536,0.033210668298933245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,12288,0.01847822301917606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,512,0.02863111098607381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,4096,0.015972443752818637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,10240,0.06154488854938083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,16384,0.9107066260443794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,7168,2.887020534939236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,32,0.009408888717492422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,2560,0.0657715532514784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,2560,0.008127999802430471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,2560,0.11483821603986953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,6144,0.1590533388985528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,6144,0.05847377909554375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,64,0.0081377774477005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,10240,0.9560026592678494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,8192,0.012279999752839407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,256,0.012879110872745514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,5120,0.012301333248615265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,65536,0.011033777561452655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,512,0.012317332956526013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,256,0.235609769821167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,7168,0.10670133431752522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,768,0.010590222146775989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,8192,0.09240977631674872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,10240,0.03090755475891961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,128,0.025103999508751765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,32,0.006160000132189856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,1536,0.19401689370473227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,2048,0.08812444739871556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,3072,0.056426664193471275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,10240,0.01408088869518704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,2048,0.06384978029463026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,512,0.007908444437715743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,128,0.0027679999669392905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,65536,0.06731644603941175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,6144,0.012302222351233164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,64,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,6144,0.11659822199079727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,5120,0.028747555282380845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,512,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,51200,0.017216889394654166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,128,0.0026888888743188647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,768,0.021253334151373968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,1536,0.008216888540320927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,4096,1.4473813374837239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,2048,0.011760888828171624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,32,0.003177777760558658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,3584,0.011801777614487542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,768,0.008207999997668797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,3072,0.0074648890230390764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,64,0.06380977895524767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,16384,1.7473093668619792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,8192,0.016008888681729633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,5120,0.013597333596812354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,2048,0.05231555634074741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,2048,0.0406328903304206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,7168,0.08788177702162002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,8192,0.050429334243138633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,10240,0.47073867585923934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,51200,0.08299289147059123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,4096,0.09345333443747626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,768,0.03967466619279649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,1024,0.00851822230550978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,51200,1.4665991465250652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,51200,0.011584888729784222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,7168,0.008221333225568136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,32,0.0038924444880750445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,16384,0.06195733282301161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,10240,0.04105066590838962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,1024,0.06169155571195814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,64,0.003071111109521654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,3584,0.008212444682916006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,3072,0.015430221954981485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,2560,0.036900444163216486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,32,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,65536,0.08513244655397202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,51200,11.524078369140625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,3584,0.0200311111079322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,65536,1.182646221584744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,4096,0.11301778422461616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,6144,0.3225919935438368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,7168,0.018935110833909776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,128,0.0033039999090962936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,6144,0.20578844017452666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,32,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,4096,0.19847910934024385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,1024,0.20975732803344727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,256,0.025926222403844196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,768,0.007345777418878343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,32,0.004893333133724001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,128,0.004514666481150521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,768,0.009399111072222391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,3072,0.16327555974324545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,10240,0.01641688909795549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,64,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,768,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,5120,0.050352887974845044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,256,0.01829333272245195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,1024,0.006123555617200003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,2560,0.005940444353553984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,12288,0.42940621905856663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,512,0.014489778214030795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,128,0.010860444770918952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,8192,0.006177777631415262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,51200,0.6160177654690212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,10240,0.01825777855184343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,512,0.04282933473587036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,3072,0.008254222571849823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,65536,0.19629244009653726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,1536,0.007104888558387756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,7168,0.03437866767247518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,1024,0.04158400164710151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,3584,0.009966221948464712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,128,0.0027591110103660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,128,0.02084355552991231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,5120,0.013367111484209696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,2048,0.021132444341977436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,64,0.46940088272094727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,768,0.009479111267460717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,1536,0.005508444375462002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,65536,2.7228605482313366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,3584,0.13149689303504095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,3584,0.012920888761679331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,2048,0.010225777824719747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,256,0.03273422188229031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,51200,0.02513155506716834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,128,0.003014222201373842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,16384,0.23627911673651802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,1024,0.010096000300513374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,3072,0.013528889252079858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,256,0.03305777907371521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,1536,0.13050044907463923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,2560,0.03312711252106561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,3584,0.010045333041085137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,256,0.00481244424978892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,5120,0.012119111087587146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,8192,0.44045154253641766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,8192,0.05293066634072197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,10240,0.031276444594065346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,512,0.006894222150246303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,10240,0.11773510773976643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,4096,0.05067822337150574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,128,0.004212444441186057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,5120,0.13546755578782824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,3072,0.8150071038140191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,12288,0.19534577263726127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,2048,0.007363555332024892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,128,0.022074666288163926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,3072,0.011736888852384357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,256,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,4096,0.012470222181744046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,10240,0.009272000028027428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,10240,0.14827911059061685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,1536,0.014670222997665405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,3072,0.01536266671286689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,12288,0.02570488883389367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,256,0.010799110763602786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,4096,0.026162667406929865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,3072,0.0177848885456721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,12288,0.01900622248649597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,256,0.0028968888024489083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,2560,0.112591995133294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,10240,0.09018399980333115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,51200,0.10363199975755479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,16384,0.06021510892444187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,1024,0.011101333631409539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,65536,0.10214489036136204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,1536,0.0879431102010939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,7168,0.4752826690673828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,16384,0.2112764385011461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,5120,0.3563226593865289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,8192,0.017470222380426195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,3072,0.13371200031704372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,2048,0.007547555698288812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,12288,0.019271110494931538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,128,0.00544888898730278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,2048,0.010270222193664974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,1024,0.09271644221411811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,12288,0.10065066814422607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,16384,0.040976888603634305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,256,0.2688222196367052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,768,0.01704266667366028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,8192,0.08737422360314263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,1024,0.027826666831970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,5120,0.04123555620511373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,8192,0.3524320125579834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,10240,0.037492444117863975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,2048,0.007639110916190677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,10240,0.007656888829337225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,65536,0.34862399101257324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,768,0.3034755653805203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,128,0.011872889267073737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,8192,0.2885884443918864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,5120,0.023746665981080797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,12288,0.03537955549028184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,7168,0.015824000040690105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,3072,0.04272355635960897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,4096,0.04201244314511617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,4096,0.012261333564917246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,65536,0.04560177856021457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,64,0.018181333939234417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,1024,0.07336444324917264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,10240,0.03727822171317206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,512,0.004866666677925321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,16384,0.023366222778956097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,8192,0.21256532933976915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,2048,0.010780444575680627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,4096,0.05916711356904772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,128,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,8192,0.5378790961371528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,65536,0.3475786579979791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,2048,0.007996444072988298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,51200,0.1101271112759908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,8192,0.013849777479966482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,12288,0.11460088358985053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,4096,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,256,0.005200888961553574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,1536,0.014696000350846184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,65536,0.2875324355231391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,2048,0.015016888578732809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,768,0.15771022107866076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,3584,1.0777084562513564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,7168,0.0240346673462126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,65536,0.06126488579644097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,6144,0.017812444104088675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,768,0.022839999861187402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,5120,0.041857779026031494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,7168,0.010199111368921068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,8192,0.017456889152526855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,2560,0.008403555386596257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,1024,0.03920000129275852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,5120,0.020635555187861126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,512,0.23427910274929473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,6144,0.04351288742489285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,16384,0.019144000278578866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,768,0.006229333165619109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,8192,0.07514488697052002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,8192,0.040715555349985756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,16384,63.359822591145836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,2560,0.01716088917520311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,8192,3.2228266398111978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,65536,0.5773342450459799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,65536,0.07438133160273235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,1024,0.07525866561465792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,3584,0.013419555293189155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,256,0.0030657777355776895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,7168,0.023178666830062866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,768,0.04807555675506592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,1024,0.026087111896938745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,8192,0.03717066513167487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,64,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,8192,0.1441440052456326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,1024,0.08474933438830906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,16384,0.028029332558314007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,10240,0.020844444632530212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,512,0.006875555548402999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,65536,0.17131111356947157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,6144,0.0354693333307902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,7168,0.11439732710520427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,65536,0.14098578029208714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,16384,0.2533351050482856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,16384,0.03602666656176249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,5120,0.03632355398601956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,3584,0.006813333266311222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,6144,0.14683822790781656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,3584,0.0251217782497406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,4096,0.13505600558386907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,128,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,512,0.00572622236278322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,6144,0.015847999188635085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,768,0.07712799972958036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,16384,0.1620559957292345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,51200,0.15430222617255315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,4096,0.7537564701504178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,3072,0.011731555892361535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,3584,0.014557333456145393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,2048,0.06545955604977079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,16384,0.025006221400366888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,1536,0.02958577871322632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,768,0.01713244451416863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,32,0.005682666682534748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,51200,0.18835555182562935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,12288,0.028191109498341877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,2560,0.2507786750793457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,3072,0.00832177781396442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,128,0.007743111087216272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,10240,6.506098853217231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,512,0.01942399972014957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,1536,0.05489066574308607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,768,0.0037137779096762338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,4096,0.018713777263959248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,7168,0.16505511601765951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,768,0.003612444632583194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,8192,0.6084026760525173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,32,0.007145778172545963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,1536,0.012107555237081317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,65536,40.06773206922743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,32,0.01893422173129188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,512,0.004912888838185204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,16384,5.206409030490451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,1536,0.04257422354486254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,51200,0.11121511459350586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,1536,0.00645955569214291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,1024,0.25465332137213814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,6144,0.014910222755538093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,256,0.011722666521867117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,10240,0.025282666087150574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,65536,0.16766933600107828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,768,0.016743999388482835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,512,0.0037582222786214617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,10240,1.566271146138509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,7168,0.02730133301681942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,16384,0.0935973326365153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,768,0.008197333249780867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,1024,0.004015111260943943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,256,0.11426399813758002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,5120,0.026458667384253606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,51200,0.14534933037228054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,7168,0.011186666786670685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,10240,0.046799110041724316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,2560,0.009602666729026372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,2560,0.01256711118751102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,3072,0.05931911203596327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,12288,0.04465244544876946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,5120,0.01793244481086731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,3072,0.006568888823191325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,256,0.0041457778877682155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,512,0.007374222079912822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,512,0.005719999886221356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,6144,0.02719288898838891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,512,0.00795288880666097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,10240,0.03267822331852383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,1024,0.04772355490260654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,2560,0.005931555396980709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,128,0.005215999980767568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,65536,0.0680826637479994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,128,0.004871110949251386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,16384,1.3617573844061956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,128,0.014880888991885714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,512,0.04424000117513868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,2560,0.10778488715489705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,1536,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,512,0.17207911279466417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,4096,0.016167999969588388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,768,0.014675555957688225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,256,0.010440000229411654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,51200,1.3916986253526475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,5120,0.2956462171342638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,10240,0.07679200172424316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,16384,0.15412533283233643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,1024,0.011689778003427716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,16384,0.031472888257768415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,12288,1.8267786237928603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,3584,0.015916445189052157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,1024,0.004929777648713854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,3584,0.032608889871173434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,128,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,3584,0.16927288638220894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,64,0.019890666007995605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,768,0.004907555464241239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,65536,0.16070132785373265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,12288,0.0395760006374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,768,0.09061155716578166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,1024,0.009194666312800514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,10240,3072,0.03311377763748169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,2560,0.16036444240146214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,64,0.0047048889100551605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,32,0.016235555211702984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,8192,0.03386844529045953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,3072,0.01594044433699714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,10240,1.7504035101996527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,768,0.007302222152551015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,51200,0.12017867300245498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,768,0.016700444949997794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,7168,0.128912885983785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,1024,0.01901333365175459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,128,0.005151111218664381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,8192,0.011595555477672152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,2560,0.012959111067983838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,8192,0.4432195557488336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,5120,0.010659555594126383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,6144,0.08301688565148248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,10240,1.496578640407986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,51200,0.6672311358981662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,2560,0.10420711172951592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,256,0.0034328889515664843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,16384,0.033879111210505165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,64,0.007584888901975419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,512,0.013247110777431063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,768,0.04900888933075798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,8192,0.20221866501702201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,768,0.04489511251449585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,64,0.002895111011134254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,5120,1.1309049394395616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,12288,0.08199110958311293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,2048,0.3085217740800646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,2560,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,10240,0.01903733279969957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,2560,0.08470222022798325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,1536,0.006621333460013072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,3584,0.014540443817774454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,7168,0.43840710322062176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,10240,0.5102160241868761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,8192,0.02583733366595374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,1024,0.00574933323595259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,8192,0.029838220940695867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,5120,0.022063111265500385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,51200,2.6466782887776694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,8192,0.01642311116059621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,256,0.012289777398109436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,512,0.015182221929232279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,7168,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,16384,0.0540231102042728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,128,0.0099653336736891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,2560,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,6144,0.02400355537732442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,7168,0.024086222052574158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,512,0.009468444519572789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,6144,0.010328888893127441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,1536,0.8469191127353244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,1024,0.023549333214759827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,512,0.004429333325889376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,51200,0.06149511204825508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,768,0.00977422214216656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,12288,0.04011822077963088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,1024,0.009956444303194681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,7168,0.020799999435742695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,12288,0.17627288235558403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,256,0.004594666676388847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,768,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,5120,0.027804444233576458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,64,0.008466666771305932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,12288,0.01464000013139513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,3072,0.020450666546821594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,3072,0.05911022424697876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,6144,1.0624844233194988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,65536,0.043057776159710355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,3072,0.05709422296947903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,512,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,12288,0.01775644388463762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,5120,0.03529599971241421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,6144,0.12514132923550075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,7168,0.1058737768067254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,3584,0.012315555579132505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,1536,0.016682666209008958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,51200,1.3365137312147353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,51200,1.730199072096083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,1024,0.009334222310119206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,65536,0.06776355372534858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,512,0.021712889273961384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,16384,0.2628568808237712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,512,0.023919999599456787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,4096,0.01071200023094813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,4096,0.008268444074524773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,65536,0.014378666877746582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,5120,0.14569689167870417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,16384,0.013922666509946188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,2048,0.014073777529928418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,8192,0.04509955644607544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,65536,0.09931466976801555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,6144,1.1296799977620442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,2048,0.006617777877383762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,1536,0.005757333503829108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,1536,0.039615111218558416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,512,0.0041680000722408295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,65536,2.932994630601671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,3584,0.055496888028250806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,51200,0.06154844495985243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,2048,0.007784888976150089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,2560,0.17328977584838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,1024,0.004390222330888112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,4096,0.034825778669781156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,5120,0.05703377723693848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,256,0.006110222389300664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,64,0.005103999955786599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,7168,0.041208889749315046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,16384,0.2471191088358561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,1536,0.022693332698610093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,2048,0.15083911683824328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,256,0.006321777900060018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,1536,0.006312888943486744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,10240,0.14624533388349745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,8192,0.15094044473436144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,7168,0.395668453640408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,16384,0.1636186705695258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,2560,0.010102222363154093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,512,0.012753778033786349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,3584,0.017154667112562392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,1536,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,1536,0.04845155609978569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,6144,0.019515555765893724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,65536,1.7220435672336156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,5120,0.07505066527260675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,7168,0.03279022375742594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,2560,0.049885332584381104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,512,0.0058871110280354815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,10240,0.32885511716206867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,2560,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,12288,0.07057955529954699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,3072,0.08611644638909234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,4096,0.013588444226317935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,65536,0.06862311230765449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,12288,0.4576631122165256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,512,0.013377777404255338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,2560,0.013623110949993134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,768,0.004296888907750447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,65536,2.9906959533691406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,5120,0.41183111402723527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,8192,0.054458667834599815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,8192,0.5546515782674154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,5120,0.08398133516311646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,32,0.005594666633341048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,8192,0.02369600037733714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,32,0.004053333153327306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,51200,0.12592621644337973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,7168,0.283735990524292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,4096,0.1315564446979099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,2048,0.009132444030708736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,768,0.02362577782736884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,3584,0.12035022841559516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,5120,0.028773334291246202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,128,0.004452444612979889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,12288,0.4077591101328532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,1536,0.01590577761332194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,65536,1.500437312655979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,7168,0.06216622061199612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,768,0.0034346667428811393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,512,0.003286222202910317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,1536,0.015608888533380298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,64,0.0030133333057165146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,10240,0.39376089307996964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,5120,0.01665599975321028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,12288,0.5903466542561849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,8192,0.013458666702111563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,512,0.006248888870080312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,51200,0.8157075775994195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,64,0.03887110948562622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,8192,3.269736819797092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,64,0.011000889043013254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,1024,0.0060159998635451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,16384,1.6578346888224285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,2048,0.13135733869340685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,4096,0.021239999267790053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3072,3584,0.011759111450778114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,512,0.08581599924299453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,2560,0.03613155417972141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,3072,5.030767228868273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,2560,0.04112711217668321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,3072,0.04642488890224033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,256,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,512,0.01643822259373135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,4096,0.011584888729784222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,1536,0.0063564446237352155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,1024,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,65536,0.16101688808865017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,3072,0.043603556023703686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,8192,0.015012444721327888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,2048,0.09605155388514201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,12288,0.32802844047546387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,1024,0.014518222875065274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,16384,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,5120,0.02344177828894721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,512,0.011777777638700275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,768,0.05889155467351278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,7168,0.013799111048380533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,6144,0.015503999259736804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,32,0.0059279998143513995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,512,0.0119991112086508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,7168,0.04760266674889458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,16384,0.6446595721774632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,3584,0.014512888259357877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,2560,0.0075502221783002215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,768,0.009367110828558603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,65536,0.10961777634090847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,768,0.00793422261873881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,2048,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,256,0.00286666676402092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,4096,0.01904800037542979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,16384,2.658156500922309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,51200,0.1429333289464315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,3584,0.11739110946655273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,10240,0.029457777738571167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,1024,0.011407111254003314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,8192,0.09892888863881429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,1024,0.07549244165420532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,256,0.0058408888677755995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,3072,0.9365777969360352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,3072,0.013758222262064615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,6144,0.017376888129446242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,512,0.0076453329788313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,16384,0.024455111887719896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,10240,0.1798106696870592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,10240,0.07351377937528822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,2560,0.021698666943444148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,1536,0.008080888953473832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,2560,0.038345777326160006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,6144,0.1538239982393053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,512,0.024458666642506916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,512,0.003265777809752358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,6144,0.30258221096462673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,51200,1.7237502204047308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,6144,0.04929244518280029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,7168,0.6124933560689291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,128,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,256,0.006520888871616787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,7168,0.0958799984720018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,16384,0.3103128804100884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,51200,0.06813066535525851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,4096,0.04649600055482653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,4096,0.015338665909237333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,128,0.007287110719415877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,65536,2.5790532430013022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,7168,0.6081368658277724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,128,0.04263110955556234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,2560,0.010991111397743225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,1536,0.011587555209795633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,16384,0.2230515480041504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,64,0.02166844407717387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,3584,0.010837333069907295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,128,0.002991111121243901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,64,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,5120,0.02304711110062069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,51200,0.05872088670730591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,3584,0.06162133481767443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,1536,0.026893332600593567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,4096,0.19445244471232095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,10240,0.04253155655331082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,10240,0.041711111863454185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,256,0.0030248889492617715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,512,0.03469333383772109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,3072,0.008126222425036961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,512,0.1434497833251953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,256,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,8192,0.01874755488501655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,32,0.017101332545280457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,128,0.002653333254986339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,3072,0.010116444693671333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,32,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,10240,0.06727911366356744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,256,0.007551111280918121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,6144,0.012720888687504662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,6144,0.014819555812411837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,2560,0.39052621523539227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,3584,0.06573600239223905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,1024,0.07343822055392794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,2560,0.014843554960356818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,2048,0.07441600163777669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,256,0.1000337799390157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,10240,0.006075555665625467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,7168,6.4799406263563375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,3584,0.8994346194797092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,2560,0.17623378170861137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,1024,0.005856888989607493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,128,0.0034222222036785553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,65536,0.07570133606592815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,12288,0.06760444243748982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,1536,0.0418817765182919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,6144,0.37010576989915633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,16384,0.49562931060791016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,256,0.02496266696188185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,16384,0.030622220701641504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,128,0.005081777771313985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,4096,0.009903999666372934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,6144,0.11587466133965386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,128,0.004032000071472592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,1024,0.004025777594910728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,8192,1.4616915384928386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,768,0.044567998912599355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,3072,0.5293546782599555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,3072,0.013979555832015144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,768,0.004121777911980947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,128,0.05280177791913351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,128,0.008285333712895712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,2048,0.02555911077393426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,512,0.0041911109454101985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,1024,0.01201600001917945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,2560,0.01124622258875105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,65536,20.789094712999134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,12288,1.5775849024454753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,1024,0.007699555820888943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,65536,0.1273626618915134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,256,0.008071111308203803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,2048,0.01051288925939136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,512,0.012700444294346703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,64,0.006934222247865465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,2560,0.010064000056849586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,5120,0.03071999881002638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,8192,0.023904000719388325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,3584,2.4681875440809464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,2048,0.03321689036157396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,5120,0.3677813212076823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,65536,0.1487493382559882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,6144,0.1339084439807468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,4096,0.020592888196309406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,5120,0.018739556272824604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,1024,0.034098668230904475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,768,0.006642666541867786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,5120,0.01571999986966451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,512,0.005933333188295364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,512,0.011796444654464722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,10240,0.24827112091912162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,2560,0.009495110975371467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,3072,0.21246933937072754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,512,0.011665778027640449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,128,0.04429866539107429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,7168,0.024308444725142583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,7168,0.038183109627829656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,2560,0.054845333099365234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,1024,0.013960889644092984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,65536,0.11887289418114556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,1024,0.4636738035413954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,16384,0.024175999893082514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,65536,9.064678615993923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,12288,0.6016977628072103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,32,0.00758844448460473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,768,0.018711111611790128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,1536,0.015836444165971544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,768,0.006956444846259222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,65536,0.07965599828296237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,512,0.003640000190999773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,1024,0.36615289582146543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,16384,0.009787555370065901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,8192,0.07065066364076403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,5120,0.012944888737466602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,6144,0.2188488907284207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,8192,0.021192888418833416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,32,0.02460799945725335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,10240,0.018394667241308425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,512,0.007792889244026608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,10240,0.05809422334035238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,1536,0.0075502221783002215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,1536,0.505624877081977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,6144,16.05447726779514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,8192,0.0203493336836497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,16384,0.09523644712236191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,16384,0.02281155520015293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,8192,0.041448887851503156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,51200,1.3261644575330946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,4096,1.664967960781521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,32,0.005148444324731827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,4096,0.010934222075674268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,768,0.006140444427728653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,128,0.012602667013804117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,5120,0.03831911087036133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,12288,0.9322035047743055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,768,0.020247111717859905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,6144,0.04747111267513699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,10240,0.05401244428422716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,128,0.0064222220745351576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,3584,0.790080017513699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,512,0.0058311112225055695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,1024,0.006335111127959357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,2048,0.00793688909875022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,4096,0.08268977536095513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,3072,0.12257066037919785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,65536,0.8918675316704644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,51200,0.08638133605321248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,1536,0.015563555889659457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,5120,0.024596444434589807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,3072,0.14519288804796007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,2048,0.040694223509894475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,65536,0.26924000846015084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,3072,0.021409778131379023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,3072,0.02516088883082072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,32,0.04996000064743889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,8192,0.19740266270107695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,2560,0.007696888513035244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,16384,0.6325199868943956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,32,0.005186666631036335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,512,0.0173333336909612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,4096,1.9792986975775824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,1024,0.007854222423500484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,65536,0.09068622191747029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,1536,0.019333332777023315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,12288,5.09712176852756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,12288,0.648180431789822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,1024,0.007537778053018782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,16384,0.29034490055508083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,1536,0.01536977787812551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,8192,0.31886310047573513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,7168,0.04507911205291748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,5120,0.02685866587691837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,16384,0.03216800093650818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,6144,0.01847199930085076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,1536,0.006325333482689328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,768,0.050586667325761586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,16384,0.6466355323791504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,32,0.00675022229552269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,4096,0.04398933384153578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,10240,4.511830223931207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,4096,0.014567111929257711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,768,0.005685333162546158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,128,0.002728888971938027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,8192,0.021389333738221064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,256,0.0041857775714662345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,3584,0.020986666282018025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,5120,0.018788443671332467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,128,0.011512888802422417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,256,0.006502222269773483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,3584,0.44569868511623806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,768,0.04540444413820902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,4096,0.006561777657932705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,1536,0.04106311003367106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,2560,1.6913831498887804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,1536,0.01553600033124288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,512,0.022272888157102797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,7168,0.018650665879249573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,4096,0.010297777752081553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,4096,0.07332799832026164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,32,0.008590222232871586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,8192,0.0890115565723843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,51200,0.6543173260158962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,3072,0.01980177727010515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,5120,0.10441244310802883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,1024,0.007839110990365347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,6144,0.05014311273892721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,3072,0.18800711631774902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,64,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,7168,0.0813466641638014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,1024,0.015087111128701104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,8192,0.020963556236690942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,768,0.030060443613264296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,768,0.030621333254708186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,1536,0.035565333233939275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,2560,0.00797155582242542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,12288,0.03195110956827799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,1024,0.006176888942718506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,3584,0.02217866645918952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,3072,0.024518221616744995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,2560,0.047784000635147095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,2560,0.06035911374621921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,1536,0.033407999409569636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,8192,0.028353777196672227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,64,0.002593777866827117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,10240,3.170426686604818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,6144,0.28894933064778644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,8192,0.8305724461873373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,64,0.06935289171006945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,1024,0.011039111349317761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,65536,0.1776862276924981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,7168,0.01330311099688212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,1024,0.014304000470373364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,8192,0.10120177931255764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,512,0.02257866660753886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,16384,2.8758826785617404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,768,0.027012444204754297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,1536,0.006335111127959357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,8192,2.5203314887152777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,12288,1.1853680080837674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,4096,0.01034755590889189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,51200,0.10724444521798028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,1536,0.007891555627187094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,12288,0.17055377695295545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,1536,0.010093332992659675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,512,0.035671111610200666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,51200,0.13737421565585667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,768,0.3177928924560547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,10240,0.4572800000508626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,3072,0.06740622387992011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,2048,0.007333333293596904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,64,0.005542222410440445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,1024,0.11047555340660943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,32,0.011181332998805575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,32,0.002620444529586368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,768,0.021198223034540813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,2560,0.012784000072214337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,1536,0.011049778097205691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,1536,0.06888000170389812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,8192,0.01370844410525428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,32,0.002814222127199173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,1024,0.027239999837345544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,128,0.005116444494989184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,768,0.020950223008791607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,4096,0.05962666538026598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,12288,0.20963199933369955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,1024,0.00645955569214291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,512,0.009892444643709395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,10240,0.10079377889633179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,8192,0.013883555928866068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,1536,0.004873777843183941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,16384,1.9051280551486547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,2560,0.1924746698803372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,4096,0.01628533336851332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,256,0.021937777598698933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,16384,0.0068897778789202375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,10240,0.4799857669406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,8192,0.007376000285148621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,32,0.010257778068383535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,256,0.005016889009210798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,12288,0.05614488654666477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,1536,0.011696889168686338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,51200,0.06961511241065131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,5120,0.13580532868703207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,2560,0.008745777938101027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,32,0.00443822228246265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,7168,0.020432000358899433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,256,0.01586577792962392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,65536,0.17025600539313424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,32,0.02442755632930332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,51200,21.46292283799913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,1024,0.3942106564839681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,10240,0.01983022193113963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,3584,0.011804444094498953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,32,0.005477333234416113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,64,0.013863999810483722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,1024,0.01129155523247189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,768,0.015498666299713982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,3072,0.011574221981896294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,2048,0.0943288869327969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,4096,0.009745777481132084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,8192,0.05026844475004408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,64,0.005616888817813661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,4096,0.016685333516862657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,3584,0.09062400129106309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,1536,0.009366222553782994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,6144,0.013175110850069257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,512,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,51200,0.49198754628499347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,32,0.009602666729026372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,65536,2.7961983150906033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,1024,0.0074142225914531285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,2048,0.029057777590221826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,2048,0.010674667027261523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,512,0.022641777992248535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,2048,0.0506533318095737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,6144,0.18361688984764946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,1024,0.02674399978584713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,2560,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,1536,0.011782222323947482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,4096,0.025955556167496577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,1024,0.08503822485605876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,5120,0.009546666509575313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,3072,0.01109955542617374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,512,0.03545066714286804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,8192,0.011789333489206104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,7168,0.04261777798334757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,65536,0.25643467903137207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,7168,0.030583110120561387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,12288,0.01665599975321028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,256,0.018509333332379658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,7168,0.018795554836591084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,51200,1.1076293521457248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,3072,0.010806221928861407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,512,0.03448266784350077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,128,0.007076444725195567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,16384,2.627643585205078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,768,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,5120,0.08456355333328247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,64,0.007836444510353936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,128,0.005067555440796746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,4096,0.08787288930681016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,32,0.03638666537072923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,7168,0.020473778247833252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,5120,0.1107306679089864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,1536,0.054104887776904635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,32,0.003109333415826162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,3584,0.00793422261873881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,3584,0.010667555862002902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,65536,0.06267111168967353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,3584,0.03186933199564616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,8192,0.012881778180599213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,768,0.003674666500753827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,3584,0.008683555656009251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,65536,0.49665599399142796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,8192,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,3072,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,5120,0.03229866756333245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,256,0.00498844434817632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,512,0.049267556932237416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,65536,0.11225244734022354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,128,0.0033608888172441055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,6144,0.0669777790705363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,512,0.10178399748272365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,128,0.007518221934636434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,4096,0.01482399966981676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,2048,0.013855111267831592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,3584,0.9824515448676215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,65536,0.01591199967596266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,7168,0.6727555592854818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,3584,0.008912888665994009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,2048,0.12134310934278701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,2048,0.5126230981614854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,5120,0.010673777924643623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,4096,0.8094239764743381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,2560,0.015677332878112793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,3072,0.04613599843449063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,6144,0.037307557132509016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,51200,0.38395200835333926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,6144,0.06622133652369182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,128,0.013460444079505073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,8192,0.02377066678471035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,5120,0.025587555434968736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,7168,0.10456622309154934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,128,0.026311111119058397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,7168,0.020005333754751418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,3072,0.010144000252087912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,51200,7.112223307291667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,2560,0.008952889177534316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,7168,0.031077331966824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,128,0.006275555739800136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,1024,0.01570666664176517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,8192,0.11606844266255696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,8192,0.14574755562676325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,64,0.003032888803217146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,6144,0.639087994893392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,128,0.005115555392371284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,51200,0.9033386442396375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,512,0.14102666907840306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,512,0.0062773335311147906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,65536,0.17140977912478975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,2560,0.01464355488618215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,512,0.003693333516518275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,4096,0.02021422154373593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,512,0.2539893256293403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,128,0.012972444295883179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,768,0.21889155440860322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,128,0.0034008889148632684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,3584,0.054208887947930225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,1536,0.052431109878751964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,768,0.02014044423898061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,10240,0.022720000810093347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,5120,0.009669333696365356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,768,0.006262222097979651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,10240,0.01831377711560991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,512,0.007932444413503012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,512,0.005535999933878581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,64,0.0028719999310043124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,512,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,3072,0.007256000406212277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,256,0.01218577805492613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,2048,0.013009777499569787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,64,0.002903999967707528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,256,0.008036444584528605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,768,0.006125333408514659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,2048,0.06259110901090834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,7168,12288,0.5401004685295953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,5120,0.15939288669162327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,6144,1.606960932413737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,6144,0.03820800118976169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,64,0.00792711145348019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,64,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,7168,0.03379999929004245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,512,0.13428444332546657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,6144,0.027484445108307734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,7168,0.02791555391417609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,2048,0.7227190865410699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,1536,0.019349333312776353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,65536,0.09999022218916152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,2560,0.01037333326207267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,65536,2.8843992021348743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,6144,0.01296533313062456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,12288,11.079069349500868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,7168,0.03312888741493225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,2048,0.04050933321317037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,1536,0.007607111500369177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,64,0.2465519905090332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,4096,0.01623644431432088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,12288,0.023397333092159692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,1024,0.006548444430033366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,7168,0.007446222007274628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,65536,0.05549155672391256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,65536,0.42459289232889813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,256,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,51200,0.19887022177378336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,6144,0.025869333081775244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,51200,0.5244933234320747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,7168,0.659447087181939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,10240,0.12610577212439641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,7168,0.08623288737403022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,65536,0.5801822344462076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,768,0.10461688703960842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,51200,0.30938755141364205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,10240,0.8743377791510688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,12288,0.03204355637232462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,65536,0.2256115542517768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,16384,0.29367465443081325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,8192,0.007975999679830339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,6144,0.12201421790652806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,6144,0.08102489180035062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,1024,0.0063928887248039246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,12288,0.1282471153471205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,32,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,6144,0.018224888377719455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,4096,0.0743502246008979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,3072,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,2048,0.01421955558988783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,2048,0.006413333531883027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,65536,0.40518755382961696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,32,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,2048,0.04116888840993246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,256,0.033674667278925575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,2560,0.3299902280171712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,16384,0.022368000613318548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,6144,0.006286222073766921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,256,0.003152888889114062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,768,0.005483555710977978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,6144,0.29400356610616046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,12288,0.16370044814215765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,1024,0.004498666773239772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,10240,0.028661333852344092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,2048,0.006719111154476802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,256,0.006120000034570694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,128,0.004625777817434735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,12288,0.040087110466427274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,4096,0.16037689314948186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,256,0.01277511070171992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,6144,0.020647111866209242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,1024,0.008812444077597724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,256,0.0030986666679382324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,8192,0.021736888421906367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,4096,0.02894044584698147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,6144,0.021108445194032457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,8192,0.018200889229774475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,3584,0.007948444121413762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,256,0.016736888223224215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,3584,0.38543023003472227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,512,0.11339110798305935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,768,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,32,0.0030248889492617715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,2048,0.006299555715587404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,6144,0.01662577854262458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,2048,0.023403555154800415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,256,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,16384,0.09434488746854995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,3072,0.013186666700575085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,65536,0.24866755803426108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,3072,0.11411555608113606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2560,4096,0.047853334082497485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,10240,0.6606604258219401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,3584,0.030093332131703694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,7168,0.03496177660094367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,8192,1.6465270784166124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,4096,0.03532622257868449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,1536,0.007802666889296637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,6144,0.010855110983053843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,8192,0.1579128901163737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,768,0.004570666700601578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,6144,0.009370666411187913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,6144,0.25716622670491535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,1536,0.013826666606797112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,768,0.0036435553597079385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,6144,0.023269333773189124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,3584,0.010393777655230628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,512,0.0032808888289663526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,8192,0.06599289178848267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,512,0.005920889063013925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,65536,0.08849955267376369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,3072,0.13274399439493814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,1024,0.04560533497068617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,32,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,65536,0.012858666479587555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,16384,0.07443644603093465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,3584,0.010392000277837118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,7168,0.015086222026083203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,5120,0.014150222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,10240,0.13109066751268175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,128,0.004310222135649787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,65536,11.970021565755209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,256,0.014704888065656027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,65536,0.07518844472037421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,7168,1.1301031112670898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,3072,0.008608889248636033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,2048,0.015015999476114908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,5120,0.2188275522655911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,3072,0.019377777973810833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,10240,0.058119999037848584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,10240,0.14747733540005153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,12288,0.07373422384262085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,16384,0.32306933403015137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,768,0.0038275557259718576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,64,0.009673777553770278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,65536,0.27270934316847056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,768,0.012677333421177335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,65536,0.013762666947311826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,3072,0.021771555145581562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,3072,0.4108702341715495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,7168,0.046557333734300405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,512,0.005767999837795894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,1536,0.04803555541568332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,6144,0.055626663896772593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1024,51200,0.053018665975994535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,6144,0.039966222312715315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,51200,0.824727111392551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,7168,0.21603733963436553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,7168,0.13091554906633165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,768,0.005121777868933148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,65536,0.10666666428248088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,16384,0.9640568627251519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,3072,0.055396444267696805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,16384,0.02808800008561876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,1536,0.11040000120798747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,5120,0.025217778152889673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,4096,0.016650666793187458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,1024,0.008227555288208855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,4096,0.942387580871582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,64,0.13902044296264648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,7168,0.00982844415638182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,65536,1.3192862404717338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,6144,0.01991555591424306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,51200,0.23811999956766763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,6144,0.04133955637613932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,512,0.00350222239891688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,16384,0.669015990363227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,1024,0.02940444482697381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,2560,0.016152888536453247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,3584,0.0173999998304579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,5120,0.6949537595113119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,512,0.006079111248254776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,2560,0.010638222098350525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,3584,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,6144,0.2219768895043267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,4096,0.01575911045074463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,10240,3.7441228230794272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,32,0.013658666776286231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,3072,0.010564444793595208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,2048,0.05602044529385037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,6144,0.5439457893371582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,768,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,3072,1.1760755115085177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,3584,0.02846933404604594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,16384,0.5885955492655436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,12288,0.12046578195359971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,6144,0.015772443678643968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,10240,0.04634933339224922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,65536,0.24984444512261283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,65536,0.8707600169711642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,5120,0.10991999838087295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,128,0.028765334023369685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,51200,0.016915554801623028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,12288,0.03645599881807963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,3584,0.038210666841930814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,7168,0.023731556203630235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,1024,0.0323884454038408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,10240,0.25744978586832684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,8192,0.01851022243499756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,2560,0.17672710948520234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,12288,0.19339999887678358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,12288,0.04781422350141737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,16384,0.5583510928683811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,10240,0.18258489502800834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,32,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,2560,0.03755911191304525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,768,0.013041777743233575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,12288,80.74940999348958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,51200,0.08321422338485718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,768,0.010461333725187512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,768,0.010982222027248807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,128,0.03466577662361993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,32,0.01277955538696713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,3584,0.009756444229020013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,768,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,768,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,128,0.018238221605618794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,51200,0.09993955824110244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,1536,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,1024,0.006729777902364731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,8192,0.011672889192899069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,1536,2.2301627265082464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,16384,0.15692888365851507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,12288,0.024075556132528517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,8192,0.03054844339688619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,32,0.005149333427349727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,6144,0.02293155590693156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,12288,0.010965333216720156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,1024,0.01570666664176517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,512,0.008314666648705801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,3584,0.1865146689944797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,12288,7.7075288560655375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,512,0.008061333662933773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,6144,0.009993777506881291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,2560,0.016664889123704698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,51200,5.122399224175347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,6144,3.7864540947808156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,32,0.044645332627826266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,3584,0.05934133132298788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,51200,0.8925608528984917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,1024,0.012703999876976013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,4096,0.39443643887837726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,256,0.0203484445810318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,128,0.00443733317984475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,256,0.012713777522246042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,8192,1.093232896592882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,64,0.003154666680428717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,5120,0.3405200110541449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,12288,0.07816444502936469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,65536,0.1473484436670939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,128,0.10547644562191433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,3584,0.011141333315107556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,768,0.013923555612564087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,3584,0.013883555928866068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,768,0.005944889038801193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,3584,0.0459191103776296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,2048,0.014224888549910652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,10240,0.01724088854259915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,4096,0.031046221653620403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,2048,0.007102222078376346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,256,0.014853333433469137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,2560,0.04823644293679131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,5120,0.08106488651699491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,1536,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,2560,0.08081777890523274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,768,0.013353777428468069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,256,0.003257777748836411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,51200,0.08487200074725681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,256,0.012299555871221753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,2560,0.027119111683633592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,8192,0.04857688811090258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,10240,0.3051982190873888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,128,0.004054222255945206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,10240,0.024978665841950312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,12288,0.017507554756270517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,512,0.024499555428822834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,65536,0.2974159982469347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,1536,0.06083555354012383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,64,0.011571555501884885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,4096,0.3501182132297092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,6144,0.040607998768488564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,65536,0.3001093334621853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,12288,0.29950043890211314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,2560,0.020842666427294414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,51200,0.9817146725124783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,12288,0.1184497806761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,6144,0.014534221755133735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,4096,0.010384000009960597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,1536,0.007074666519959767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,6144,0.08506400055355495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,12288,0.09888622495863174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,65536,0.08862488799624973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,6144,0.010956444674068026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,4096,0.01475644442770216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,3072,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,2560,0.006177777631415262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,5120,0.011647111011876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,4096,0.10440978076722886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,128,0.002899555489420891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,4096,0.008423999779754216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,2048,0.052095111873414784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,3072,0.02218399941921234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,512,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,6144,0.018787556224399142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,3584,0.009908444351620143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,16384,0.009918221996890174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,4096,0.023598222268952265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,128,0.025205334027608235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,2560,0.008995555341243744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,65536,0.17020977867974174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,51200,0.1833395560582479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,2560,0.2405688762664795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,10240,0.717650678422716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,16384,0.47720977995130753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,10240,0.07332799832026164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,5120,0.011391999820868174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,3072,0.02293511066171858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,3584,0.021614222062958613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,12288,0.04061155517896017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,16384,0.6553449100918239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,7168,0.21748977237277559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,64,0.04633511106173197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,4096,0.10696444246504043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,1024,0.05412711037529839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,51200,0.22723023096720377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,10240,0.3901324537065294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,12288,0.2542311085595025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,2560,0.017869333426157635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,512,0.005746666755941179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,16384,0.45780711703830296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,768,0.0059004442559348206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,16384,4.7975137498643665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,10240,0.4493128988477919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,65536,0.19080889225006104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,256,0.004940444396601783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,5120,0.013268444273206921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,3584,0.01793599956565433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,8192,0.3457324504852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,2048,0.03649866580963135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,51200,0.11346399784088135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,16384,0.03686044282383389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,7168,0.34041510687934023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,128,0.1845386690563626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,6144,0.1011013322406345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,256,0.13662577999962702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,512,0.007903111477692923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,1024,0.2318497763739692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,12288,0.10397955444124009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,64,0.00618577789929178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,3072,0.013674666484196981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,512,0.005113777601056629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,65536,1.5174782011244032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,128,0.12359733051723903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,1536,0.022816888160175745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,2560,0.08078489038679335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,10240,0.12735733720991346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,2560,0.024851555625597637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,51200,4.230662027994792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,8192,0.08439111047320896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,10240,0.05849155452516344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,3072,0.030015110969543457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,4096,0.02351733379893833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,256,0.009102221992280748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,256,0.005655111124118169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,16384,0.06291821930143568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,1536,0.009727111293209923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,3072,0.009840889109505547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,2048,0.019211555520693462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,4096,0.010760000182522668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,4096,0.015612444943851896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,32,0.0026773332307736077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,6144,0.01885333326127794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,2560,0.0613608890109592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,5120,0.0723866687880622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,65536,0.14383999506632486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,3584,0.006780444333950679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,512,0.008720888859695857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,256,0.02368444369898902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,64,0.011839110818174152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,8192,0.04900177650981479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,32,0.007188444336255391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,51200,0.5385680198669434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,51200,0.2060613367292616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,512,0.006778666542636023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,1024,0.010420444111029306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,2048,0.007344000041484833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,512,0.014056889547242058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,64,0.0053431110249625305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,12288,0.07975288894441393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,3072,0.06213066975275675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,51200,0.71524445215861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,256,0.00282222218811512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,512,0.03565511107444763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,128,0.007009777757856581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,10240,0.022081777453422546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,256,0.008645333349704742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,3584,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,64,0.002769777758253945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,7168,0.1060568888982137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,12288,0.12454222308264838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,7168,0.09285955296622382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,10240,0.1430071062511868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,12288,0.04868977930810717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,12288,5.721179538302952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,3072,0.023656888140572443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,7168,0.026778666509522334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,768,0.025408888856569927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,65536,1.3275440004136827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,256,0.004335999902751711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,3072,0.0125946667459276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,4096,3.5135298834906687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,128,0.005331555588377847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,12288,0.09525066614151001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,4096,0.008346666892369589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,51200,2.2003705766465926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,4096,0.017608889275126986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,65536,0.9938693576388888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,2560,0.05145333210627238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,3584,0.15568355719248453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,1024,0.00902844468752543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,4096,0.01812266641192966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,12288,0.03539822167820401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,10240,0.19440621799892852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,4096,0.036116444402270846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,12288,0.4689760208129883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,5120,0.015616888801256815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,8192,0.1476853291193644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,5120,0.02839822239345974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,4096,0.0323884454038408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,256,0.0031795555518733132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,65536,0.0684079991446601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,16384,0.02368800010946062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,1536,0.0735751125547621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,3072,0.0074924445814556545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,64,0.010668444136778513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,128,0.05101244317160713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,8192,0.05409155620468987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,128,0.01124622258875105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,256,0.008581333690219456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,1536,0.122670226626926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,7168,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,10240,0.09242222044203018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,1024,0.039381331867641874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,64,0.008783111141787635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,2560,0.021458667185571458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,128,0.0028702221396896574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,65536,2.41886223687066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,10240,0.01805688937505086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,768,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,512,0.01015911085738076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,32,0.003740444365474913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,51200,5.777997334798177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,5120,0.013597333596812354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,512,0.020587555236286588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,256,0.013611555927329592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,7168,0.028958221276601154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,16384,0.024502222736676533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,64,0.032608889871173434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,4096,0.02899644441074795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,8192,0.12478221787346734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,7168,0.018587556150224473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,3584,0.06583022409015231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,3584,0.0942222211096022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,64,0.005398222141795688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,128,0.0041786668201287585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,768,0.006587555425034628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,3584,0.0595911078982883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,3072,0.012168000141779581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,10240,0.10855733023749457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,51200,0.5120186805725098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,1024,0.009504000345865885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3072,7168,0.24748889605204263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,64,0.0028613333900769553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,2560,0.5796328650580512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,4096,0.016999999682108562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,2048,0.16295733716752794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,1024,0.028986665937635634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,32,0.0026968889352348116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,512,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,2560,0.008276444342401292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,65536,0.44754309124416775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,10240,0.2698711024390327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,128,0.0035422220826148987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,12288,0.03131022387080722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,3584,0.009312888814343346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,256,0.007017778025733099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,7168,0.027143999934196472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,1536,0.005970666805903117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,5120,0.06257510847515531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,3584,0.006144000010357962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,3584,0.04297866755061679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,128,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,8192,0.24447467592027453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,7168,0.01090844472249349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,64,0.005990222096443176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,64,0.015300444430775113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,128,0.003076444483465619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,6144,0.24171556366814506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,6144,0.01275466630856196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,128,0.005611555443869696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,256,0.005283555636803309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,51200,0.8998692830403646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,5120,0.011418666276666852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,3584,0.020763556162516277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,128,0.01130488928821352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,2048,0.010211555494202508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,2560,0.10905066463682388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,4096,0.020656888683636982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,2560,0.28163909912109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,3584,0.028931554820802476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,5120,0.589446226755778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,64,0.010021333065297868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,8192,0.07487822241253324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,51200,0.10213244623608059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,256,0.003096888876623578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,1024,0.011767110890812345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,256,0.00500177757607566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,12288,0.3070444530910916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,65536,1.2997289233737521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,1024,0.004281777888536453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,16384,0.3254035578833686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,3072,0.0655075576570299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,4096,0.010480889015727572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,7168,0.02787999974356757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,5120,0.0457857780986362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,512,0.006861333217885759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,12288,0.007872888611422645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,12288,0.13514933321211073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,1024,0.0047493332790003884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,1536,0.011680000358157687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,256,0.004830222162935468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,10240,0.29613423347473145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,51200,0.1564275556140476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,4096,0.012004444168673621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,3584,0.007000000112586551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,32,0.005141333159473208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,51200,0.3268382284376356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,64,0.0025475554996066624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,1536,0.06518577867084079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,16384,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,8192,0.030427555243174236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,7168,0.01091288857989841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,256,0.004743111216359668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,3584,0.00868266655339135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,32,0.003370666669474708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,7168,0.023741333021057978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,256,0.005436444448100195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,5120,0.016639999217457242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,7168,0.3032284312778049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,3584,0.018671999375025432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,2560,0.07774488793479072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,128,0.01644622286160787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,128,0.027687999937269423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,2048,0.3594213326772054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,2560,0.006966222491529252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,16384,4.189158969455295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,4096,0.08136622111002605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,512,0.057222220632765025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,4096,0.007916444705592262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,128,0.0027351110345787476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,16384,0.2327200041876899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,5120,0.03442755672666762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,5120,0.0695306658744812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,768,0.0805600020620558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,2048,0.006646222124497096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,6144,0.022416889667510986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,16384,1.21561951107449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,1024,0.06126311090257433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,16384,0.04001599881384108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,16384,0.04208088914553324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,128,0.006245333287451003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,1536,0.013533333109484779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,6144,0.01757866640885671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,128,0.013936888840463428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,1024,0.02202755543920729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,10240,0.03192799952295091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,65536,0.15510577625698513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,3072,0.010668444136778513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,2048,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,7168,0.020976000361972388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,2048,0.005390222287840313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,768,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,768,0.005766222046481238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,64,0.00609688874748018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,2048,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,5120,0.011622221933470832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,12288,0.03028622269630432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,65536,0.13976889186435276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,65536,0.09671199984020656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,2048,0.006650666809744305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,32,0.007180444068378872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,2560,0.013807999591032663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,512,0.021482666333516438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,256,0.03196444445186191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,16384,0.07518221934636433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,16384,0.1191893286175198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,256,0.003637333297067218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,65536,0.3552435504065619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,64,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,7168,0.01421955558988783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,12288,0.08480444219377305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,2048,0.007612444460391998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,7168,0.2982470989227295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,6144,0.09276888767878215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,10240,0.02371911042266422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,512,0.139082670211792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,6144,0.14537155628204346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,2048,0.00816888858874639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,4096,0.01660977800687154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,1536,0.006396444307433234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,8192,0.0684773325920105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,512,0.0033244445092148254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,128,0.004204444587230682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,256,0.0052471111218134565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,51200,0.04739911026424832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,1024,0.004624888714816836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,1024,0.003989333493842019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,4096,0.008443555898136562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,10240,0.21023821830749512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,4096,0.00831911133395301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,5120,0.012936000194814471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,64,0.00544799988468488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,3072,0.012874667015340594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,1536,0.005793777604897817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,10240,0.14972266885969374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,6144,0.04533066683345371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,256,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,16384,0.08506311310662164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,4096,0.055824889077080615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,32,0.004115555435419083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,2048,0.01945333348380195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,64,0.21419821845160592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,256,0.002985777747299936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,3584,0.009568889107969072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,10240,0.031039112144046362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,512,0.019861333900027804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,3072,0.00758844448460473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,32,0.006682666639486949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,10240,0.029339555237028334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,65536,10240,0.5386577712164985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,3072,0.015522667103343539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,3072,0.007827555967701806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,2048,0.007277333074145847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,256,0.24093598789638945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,2048,0.00740533322095871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,32,0.011145778000354767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,2048,0.006030222194062338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,32,0.004480000171396467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,6144,0.01351288871632682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,1024,0.0758097767829895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,7168,0.018194667167133756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,256,0.03760266635153029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,4096,0.013328889177905189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,128,0.00775999989774492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,3072,0.007865777446164025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,1536,0.004451555510361989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,3584,0.008649778034951951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,32,0.0052684446175893145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,512,0.00664533343580034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,256,0.015232889188660515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,51200,24.444654676649307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,6144,0.029021332661310833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,128,0.009703999592198266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,128,0.004078222231732475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,8192,2.4846710628933377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,10240,0.07287822167078654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,512,0.009303111169073317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,1024,0.4908906618754069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,768,0.040728890233569674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,12288,0.06087911128997803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,1024,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,2560,0.009280888570679558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,65536,0.38222487767537433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,10240,0.017292444904645283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,3584,1.153309292263455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,10240,0.06791200240453084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,1024,0.005168000029193031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,65536,0.17083466053009033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,2560,0.07204533285564847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,64,0.0031111112071408164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,10240,1.9934719933403864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,5120,1.525777816772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,5120,0.08657333585951064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,1536,0.007921777665615082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,5120,0.37581690152486164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,51200,0.11219022009107803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,12288,0.02782577938503689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,5120,0.011181332998805575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,7168,0.01605422298113505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,5120,0.07864622275034587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,7168,3.0040808783637156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,3584,0.009934222532643212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,2560,0.0292231109407213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,12288,0.3172489007314046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,4096,0.05123022198677063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,3584,0.010989333192507425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,256,0.004557333472702238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,32,0.007901333272457123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,512,0.004582222137186262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,128,1.5759467018975153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,51200,0.35488178994920516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,1536,0.03341599967744615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,768,0.015487111277050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,256,0.004214222232500712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,64,0.141628450817532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,3072,0.5920026567247179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,2560,0.007176000210973952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,2048,0.04788355694876777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,16384,0.038415110773510404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,32,0.0028311111446883944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,6144,0.013577777478430005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,5120,0.12457688649495442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,65536,1.8290284474690754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,5120,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,65536,0.6706506411234537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,2048,0.018399111098713346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,6144,0.01344622257683012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,7168,0.06795110967424181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,5120,0.0129120002190272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,7168,0.6055022345648872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,12288,0.029706666866938274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,65536,0.24175198872884116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,10240,0.03475288881195916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,65536,0.22220799658033583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,4096,0.05278933379385206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,3072,0.01145155562294854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,7168,0.03501510951254103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,65536,7.488054063585069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,1024,0.018290667070282828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,2048,0.008390222158696916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,8192,0.0439937776989407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,8192,0.031559109687805176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,3584,0.014813333749771118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,3584,0.10809955332014297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,512,0.005555555638339784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,12288,16384,0.1632195578681098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,4096,0.08045155472225614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,10240,3.578096177842882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,1024,4.2836587693956165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,1536,0.0076142218377855085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,64,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,512,0.01200355589389801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,12288,0.02202044427394867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,65536,2.8506658342149525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,1536,0.0266311119000117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,4096,0.018383999665578205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,1024,0.08692710929446751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,64,0.002765333279967308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,128,0.007967111137178209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,512,0.005691555639108022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,16384,0.03861688905292087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,16384,0.35295645395914715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,256,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,10240,0.17753422260284424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,51200,0.057468447420332164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,512,0.00443822228246265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,16384,0.011648889217111798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,128,0.011320888996124268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,8192,0.034324444002575345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,4096,0.01608088943693373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,3584,1.6130364735921223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,2048,0.10153777731789483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,768,0.0047937776479456155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,256,0.03703022334310744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,8192,0.03502577874395583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,64,0.00519822206762102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,8192,0.013948443863126965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,128,0.025044444534513686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,128,0.007425777614116669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,51200,0.19448888301849365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,16384,0.10498044225904678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,512,0.003579555700222651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,2048,0.019736889335844252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,12288,0.006810666786299811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,12288,0.18945333692762586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,3072,0.016819554898473952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,2048,0.00868711123863856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,1024,0.0039013334446483185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,51200,0.1009991102748447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,1536,0.013780444032616086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,256,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,8192,0.41146667798360187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,8192,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,7168,0.017290666699409485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,3584,0.020046222541067336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,512,0.00552977787123786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,3072,0.11849955717722575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,128,0.027888000011444092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,65536,0.06954399744669597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,16384,0.6421279907226562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,2048,0.008861333131790161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,32,0.002622222320901023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,128,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,12288,0.13437422116597494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,5120,0.014505777094099255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,64,0.0052879999081293745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,128,0.005473777651786804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,2560,0.009717333647939894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,3072,0.06653066476186116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,128,0.004500444564554426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,128,0.0041555555330382455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,10240,0.02219377789232466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,64,0.002740444408522712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,768,0.0033928888539473214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,64,0.0074675555030504865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,12288,0.006242666807439592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,512,0.0037360000941488477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,7168,0.12056088447570801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,6144,0.3953182167477078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,3072,0.9496373070610894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,512,0.00963466697269016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,5120,0.013763555222087435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,7168,0.045568890041775174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,6144,0.02089244458410475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,4096,0.008255110846625434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,10240,0.053921778996785484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,512,0.00703022215101454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,51200,8.24043443467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,768,0.050943109724256724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,64,0.0032879999942249725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,3072,0.012651555240154266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,8192,0.01538755496342977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,1536,0.00850933293501536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,4096,0.01016444464524587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,2048,0.09302577707502578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,4096,0.008948444492287105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,2048,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,1024,0.012772444221708508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,65536,3.431802749633789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,10240,0.18572889433966744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,5120,0.012651555240154266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,2560,0.01719999975628323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,768,0.015111999379263984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,2560,0.12159377998775905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,7168,0.014305777019924588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,128,0.0053591111467944244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,128,0.0028657778683635923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,65536,4.184145609537761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,256,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,2560,0.027287110686302185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,51200,0.2732844352722168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,128,0.003999999827808804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,1024,0.005199111170238919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,8192,0.1963555547926161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,128,0.02293955617480808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,10240,0.034760889079835676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,51200,0.44395822948879665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,6144,0.23883022202385795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,1024,0.00886933339966668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,51200,0.053447110785378345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,2048,0.012347555822796293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,128,0.008596444295512306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,16384,0.02181244393189748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,2048,0.007530666887760162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,12288,0.021578666236665513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,768,0.0058702222175068325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,16384,0.02312177750799391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,1536,0.018375999397701688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,3584,0.010254222485754225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,7168,1.2131413353814018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,4096,1.6353609297010634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,1536,0.36927376853095156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,2048,0.06930844651328193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,7168,0.12289955880906846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,4096,0.012685333689053854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,256,0.0037493333220481873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,3584,0.01993777851263682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,256,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,65536,0.23075466685824922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,3072,0.044532444741990834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,512,0.4007999897003174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,12288,0.0816462238629659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,1536,0.007183111376232571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,512,0.09941333532333374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,2560,0.11769510640038384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,4096,0.32743999693128795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,256,0.04030488928159078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,4096,0.018544889158672757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,1024,0.03000266684426202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,10240,0.6876008775499133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,768,0.012624888784355588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,2048,0.006462222172154321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,1024,0.006424888968467712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,256,0.004671111288997862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,51200,0.9605147043863932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,2560,0.008921778036488427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,2560,0.010072000324726105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,8192,0.3996186786227756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,1024,0.005509333478079901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,256,0.006685333533419504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,512,0.012366222010718452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,1024,0.02814399864938524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,7168,0.2397235499487983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,10240,0.45256355073716903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,1024,0.1509208944108751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,512,0.003918222255176968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,3072,0.010339555641015371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,12288,0.08731733428107367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,6144,0.012275555895434486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,8192,0.036257776949140764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,16384,0.025166221790843542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,3584,0.011839110818174152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,12288,0.052339557144376964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,7168,0.13850755161709255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,3584,0.008216888540320927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,32,0.00305600009030766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,3072,0.030809776650534734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,2048,0.05572444200515747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,32,0.004287111262480418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,16384,0.030890666776233252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,64,0.011007111105653973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,768,0.0035982223020659555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,768,0.02773511078622606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,8192,0.015621332658661736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,7168,0.09397866990831162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,8192,0.08463199933369954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,10240,0.008396444221337637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,4096,0.15490489535861546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,6144,0.024101333485709295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,65536,0.13723466131422254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,1024,0.008636444807052612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,2048,0.01722755531469981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,16384,0.48844443427191836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,4096,0.011766222616036734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,65536,0.056793775823381215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,7168,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,12288,0.03413688805368211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,6144,0.05438399977154202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,8192,0.0122097780307134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,6144,0.015297777122921415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,7168,0.005984000033802456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,4096,0.35965689023335773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,8192,0.15830577744377985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,512,0.00509066672788726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,5120,0.011121778024567498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,128,0.003015999992688497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,7168,0.015000889698664347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,6144,0.013756444056828817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,5120,0.0583768884340922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,8192,0.01552177800072564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,1536,0.04159999887148539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,4096,0.009842666486899057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,2560,0.009697777529557547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,12288,0.16411644882626003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,7168,0.1654720041486952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,1024,0.01683288812637329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,7168,0.15632800261179605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,51200,0.23271555370754668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,5120,0.06501866711510552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,768,0.04853777752982246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,6144,0.06486488713158502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,128,0.0047688889834615915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,1536,0.027818666564093694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,10240,0.22607554329766166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,10240,0.18380710813734266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,256,0.005111111121045219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,2048,0.006576889091067844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,6144,4.192249721950955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,512,0.037853333685133196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,64,0.01626133256488376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,3072,0.08133955796559651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,128,0.019537778364287484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,51200,0.35314223501417374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,768,0.005140444470776452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,2560,0.008162666526105668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,64,0.0028159999185138275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,3072,0.012864888542228274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,3584,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,2560,0.014527110589875115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,1024,0.0060159998635451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,768,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,6144,0.019966221517986722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,5120,0.015263110399246216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,2048,0.013912000589900546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,5120,0.07928622431225248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,8192,0.16849777433607313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,3584,0.009094222552246517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,256,0.00563822231358952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,768,0.0038000001675552795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,8192,3.5935770670572915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,1024,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,2560,0.06022044685151842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,51200,0.060544888178507485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,256,0.030702223380406696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,12288,0.5924444198608398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,51200,0.13419822851816812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,7168,1.0615688959757488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,51200,0.15306577417585585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,6144,0.014531556102964612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,6144,0.01832177738348643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,4096,0.18183555867936876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,8192,0.013031110995345645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,51200,0.08562311198976304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,64,0.005112888912359874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,512,0.02050844497150845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,7168,0.05609066618813408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,512,0.18712088796827528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,51200,0.09886488649580215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,4096,0.04354666670163473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,12288,0.634594652387831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,768,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,8192,0.024660444921917383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,16384,0.008648888932334052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,10240,0.2335902320014106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,7168,0.44195466571384007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,512,0.00518133325709237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,512,0.005407111098368962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,8192,0.017656000124083627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,16384,0.02511111067401038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,5120,0.029975109630160864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,12288,0.06641422377692328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,10240,0.057879110177357994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,6144,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,2048,0.03475822342766656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,8192,0.019455111689037748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,10240,0.30204889509412974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,16384,0.042690667841169566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,10240,0.01831822262869941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,1536,0.007782222496138678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,1024,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,5120,0.27345776557922363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,3584,0.12671911716461182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,1536,0.006607999818192587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,2560,0.015440889530711703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,12288,0.05681866407394409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,16384,0.21075467268625894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,3072,0.007347555624114142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,1024,0.010045333041085137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,768,0.005496888938877318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,512,0.014577777849303352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,768,0.003524444583389494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,4096,0.018555555078718398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,4096,0.03105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,5120,0.26010934511820477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,7168,0.016439111696349252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,32,0.012096000214417776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,5120,12288,1.5032907062106664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,10240,0.03286666671435038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,6144,0.07285066445668538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,7168,0.06748800145255195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,2560,0.01627200014061398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,128,0.003588444242874781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,1536,0.9961421754625109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,4096,0.01089333328935835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,3584,0.027850667635599773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,6144,1.1414355172051323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,768,0.005888888819350137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,64,0.0031226666437255014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,10240,4.151419745551215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,7168,0.2555057737562391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,512,0.006682666639486949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,2048,0.0074551113777690465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,512,0.06935911046134101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,1024,0.08092444472842746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,256,0.005884444548024072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,51200,0.6693475511338977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,2048,0.02386577758524153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,3584,0.011314666933483548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,6144,0.006267555471923616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,2048,0.12773333655463323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,128,0.005745777653323279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,7168,0.01571555601225959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,3584,0.012178666889667511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,2048,0.03942133320702447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,4096,0.01442488862408532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,1536,0.02825688984658983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,16384,0.1978222264183892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,32,0.008323555191357931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,4096,0.015156444576051501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,1536,0.039072000318103366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,2048,0.005240889059172736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,128,0.006125333408514659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,10240,0.027683556079864502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,2560,0.40311556392245823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,3584,0.06341689162784152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,65536,14.594689263237846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,64,0.007612444460391998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,16384,0.05434933304786682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,12288,0.02089777754412757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,65536,0.1711519956588745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,4096,0.008722667064931657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,256,0.005560889012283749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,10240,0.21065333154466417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,4096,0.0090515555606948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,512,0.005250666704442766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,512,0.0033688888781600525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,4096,0.18805510467953154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,4096,0.017276444368892245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,4096,0.010217777556843227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,16384,0.43087021509806317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,768,0.055585775110456676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,3584,0.037323554356892906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,65536,0.9770399729410807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,1536,0.007294221884674496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,64,0.017200888858901132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,6144,0.16114666726854113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,32,0.008903111020723978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,8192,0.016777777009540133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,12288,0.04723911152945625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,51200,5.711394839816624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,64,0.052770667605929904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,512,0.003989333493842019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,2048,0.008551110823949179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,10240,0.08101066615846422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,3072,0.021607110897699993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,8192,0.030415111117892798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,10240,0.01720622181892395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,12288,0.018949333164427016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,128,0.003093333293994268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,768,0.07818222045898438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,8192,0.7061795658535428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,12288,0.39237154854668516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,5120,0.013353777428468069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,8192,0.00961599995692571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,512,0.01591111057334476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,3584,0.10529422097735935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,2560,0.34177422523498535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,16384,0.4259022341834174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,3584,0.7862559954325358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,5120,0.03086844417783949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,6144,0.17276444700029162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,2560,0.009399111072222391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,512,0.0032106666929192017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,3584,0.9039173126220703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,12288,0.8535920249091254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,64,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,3584,0.02605155607064565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,51200,0.008953777452309927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,8192,0.019853333632151287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,2048,0.007353777686754863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,768,0.004005333201752769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,1024,0.00907199995385276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,8192,0.4887235429551866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,32,0.0028444443725877335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,5120,0.020207999481095206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,10240,0.02083555526203579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,10240,0.022823111878501043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,128,0.006695111178689533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,51200,0.09419466389550103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,51200,0.08455733458201091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,32,0.005429333282841577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,3072,0.1293555498123169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,2560,0.09106666511959499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,512,0.019387554791238572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,51200,0.854697757297092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,10240,0.01386755539311303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,1536,0.008087111016114553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,2048,0.1823964383867052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,768,0.007879999776681265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,10240,0.18890133168962267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,128,0.02793955471780565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,128,0.002760888801680671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,8192,0.540246221754286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,4096,0.008805333740181392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,768,0.0517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,1536,0.06339199675454034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,12288,0.03526844580968221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,65536,1.279798189798991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,8192,0.04872266782654656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,3072,0.033187554942237005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,3584,0.00887288898229599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,64,0.010936000280910067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,768,0.005398222141795688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,2048,0.03035555614365472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,32,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,512,0.031013333135181006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,2048,0.006552000012662675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,6144,0.1070515579647488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,4096,0.15974756081899008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,12288,0.03058755397796631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,8192,0.0397093329164717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,16384,0.6885253588358561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,1024,0.1033875544865926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,2560,0.008103110724025303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,2560,0.051988446050220065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,16384,0.021925333473417494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,8192,0.13584089279174805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,16384,0.021366222037209406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,6144,0.013064000341627332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,65536,0.7774364683363172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,5120,0.04763199885686239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,3072,0.011018666956159802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,3584,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,64,0.006821333534187741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,768,0.0069022224181228215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,1024,0.021130666136741638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,256,0.002965333354141977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,768,0.07971022526423137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,128,0.018550222118695576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,5120,0.15503377384609646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,2560,4.227793799506293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,2048,0.01464355488618215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,128,0.01993777851263682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,512,0.010796444283591377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,512,0.016211556063758004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,3072,0.014256889621416727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,64,0.01697688963678148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,65536,0.4220088852776422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,3584,0.00867377801073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,3072,0.017283555534150865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,4096,0.011156444748242697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,10240,0.018062222335073683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,128,0.07295022408167522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,1024,0.030568000343110826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,64,0.007121777368916406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,512,0.016913778252071805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,1024,0.006325333482689328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,64,0.005210666606823604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,128,0.021073778470357258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,10240,0.018363555272420246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,10240,0.028167112006081477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,7168,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,64,0.005149333427349727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,128,0.003321777822242843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,128,0.005010666532648934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,6144,0.017852443787786696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,7168,1.6024640401204426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,4096,0.008045333127180735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,12288,2.6678844028049045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,3072,0.01443733274936676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,10240,0.039368887742360435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,3072,0.009233777721722921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,8192,0.16542577743530273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,3584,0.06698755423227946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,64,0.008641777767075432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,12288,0.26330044534471303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,768,0.011588444312413534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,3072,0.11576444572872585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,6144,0.008613333106040955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,256,0.007374222079912822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,5120,0.09827288654115464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,1536,0.0047884442740016514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,7168,1.9213725195990667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,12288,0.2625208960639106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,4096,1.4401199552747939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,3072,0.10051021973292033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,6144,0.21360354953342012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,768,0.33186133702596027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,10240,4.6839243570963545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,768,0.0070879997478591064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,2560,0.029486222399605647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,256,0.06388888756434123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,12288,0.2314995659722222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,12288,0.11844178040822347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,64,0.00720355576939053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,32,0.047172443734274976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,10240,0.024215110474162634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,768,0.03477422065205044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,2048,0.007120889094140794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,12288,0.04831022355291578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,6144,0.00832088871134652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,8192,0.01754044493039449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,7168,0.018799111247062683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,7168,0.105712890625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,8192,0.4202337794833713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,3072,0.06527288754781087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,1024,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,3584,0.030495110485288832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,12288,0.006034666465388403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,12288,0.06024710999594795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,1024,0.012821333275900947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,1536,0.004575110971927643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,4096,3.55095206366645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,5120,0.1775075594584147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,65536,0.1329537762535943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,6144,0.013450666434235044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,768,0.006028444402747684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,16384,1.5649644003974066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,1024,0.005015111217896144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,3584,0.11335821946461995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,32,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,8192,0.07621600230534871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,16384,0.2945724328358968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,12288,0.007518221934636434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,4096,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,2048,0.23298311233520508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,65536,0.3034711149003771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,2560,0.2316497696770562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,10240,0.0869599978129069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,4096,0.02736088964674208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,16384,0.0244835548930698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,768,0.21041866143544516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,51200,0.06856177912818061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,16384,4.212332407633464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,3072,0.030883553955290053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,1024,0.007499555746714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,1536,0.0979786647690667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,5120,0.05619733201132881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,64,0.011249778171380361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,2048,0.025677333275477093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,16384,0.06858844227261014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,32,0.011717333561844297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,7168,0.010871999793582492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,65536,2.189503139919705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,8192,0.039320889446470476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,1536,0.008363555702898238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,128,0.004407111141416761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,1024,0.0151235560576121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,10240,0.07166577710045709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,16384,0.07563822136984931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,12288,3.113292482164171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,12288,0.060512887107001416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,3584,0.008647110727098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,10240,0.5558960172865126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,3072,0.009852444132169088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,32,0.0034648889882696998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,768,0.013172444370057849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,2048,0.026857778429985046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,2048,0.038506666819254555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,2048,0.013743110828929476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,51200,1.9777751498752172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,7168,0.02923111120859782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,512,0.004939555707905027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,768,0.044657776753107704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,51200,0.38593512111239964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,3584,0.01240888900227017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,51200,0.43537065717909074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,5120,0.01310400002532535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,8192,1.0288417604234483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,16384,0.01111288865407308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,5120,0.019890666007995605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,3584,0.0069297779765393995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,12288,0.014194667339324951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,5120,0.009174221919642555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,512,0.0461733341217041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,2048,0.027694221999910142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,12288,0.2380088965098063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,2048,0.010286221901575724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,2560,0.2883404360877143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,128,0.009210666848553551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,4096,0.01815555493036906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,128,0.020145777199003432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,128,0.015041778484980265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,65536,0.12212267186906602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,1024,0.03698577814631992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,256,0.0053342220683892565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,6144,0.018041777941915724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,65536,0.11913511488172744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,32,0.0027102221631341507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,32,0.020807999703619216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,2048,0.04212444358401828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,512,0.019966221517986722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,12288,0.015784000356992085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,51200,0.05739555756251017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,8192,0.15950666533576116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,128,0.005418666534953647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,2048,0.006992888947327931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,3584,0.02147644427087572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,7168,0.05383200115627713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,10240,0.016875555117925007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,256,0.0027520000520679685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,7168,0.011001778145631155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,10240,0.168250666724311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,1536,0.01016088906261656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,512,0.007640000018808577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,1536,0.011405333048767514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,1024,0.007768889268239339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,3072,1.2778150770399306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,128,0.00599644457300504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,3584,0.00996888925631841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,8192,0.028823110792371962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,6144,0.015202666322390238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,51200,6.444135030110677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,32,0.05902489026387533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,768,0.004907555464241239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,64,0.0035342222286595237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,51200,1.5187884436713324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,4096,0.9195857577853732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,16384,0.07315911187065972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,51200,0.05794399976730347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,768,0.014733332726690503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,2048,0.015311999453438653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,768,0.010690666735172272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,64,0.0031271111220121384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,65536,1.362203598022461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,10240,0.02386933399571313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,3584,0.015370666980743408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,768,0.006035555568006303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,64,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,2048,0.006138666636413998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1536,2560,0.008396444221337637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,10240,0.08733066585328843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,65536,0.07012799713346693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,128,0.012190221912331052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,10240,0.02203377750184801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,32,0.0033617777129014335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,64,0.00775111135509279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,2560,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,8192,0.12264977561102973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,1024,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,51200,0.20304355356428358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,1536,0.038182222180896334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,512,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,3072,0.10144355561998154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,1024,0.003970666478077571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,51200,0.26871199078030056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,3584,0.021047110358874004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,2560,0.006229333165619109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,512,0.011340444286664328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,16384,0.0901599989997016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,64,0.002817777709828483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,4096,0.009973333113723332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,7168,0.2613102330101861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,5120,0.15685421890682644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,128,0.004631999880075455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,64,0.011506666739781698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,7168,0.015142222245534262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,3072,0.02198666665289137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,8192,0.015684444043371413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,51200,3.4076258341471353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,12288,0.2554551230536567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,3072,0.0835226641760932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,10240,0.03223022156291538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,12288,0.19241510497199166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,1536,0.014935111006100973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,16384,0.0063866666621632045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,5120,0.028688000308142766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,4096,0.026628444592158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,7168,0.024078221784697637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,32,0.1078453328874376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,64,0.0052488889131281115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,10240,0.007084444165229797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,768,0.2716817855834961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,7168,0.049244443575541176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,16384,0.31414487626817494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,4096,0.025053333905008104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,7168,0.15789333979288736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,16384,0.023080888721677992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,6144,0.17211022641923693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,1536,0.007846222155623967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,7168,0.01089422239197625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,3072,0.006084444622198741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,512,0.007297777467303806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,6144,0.006075555665625467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2560,1024,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,6144,0.1642355521519979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,10240,0.13331111272176108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,3584,0.10283288690778945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,3584,0.04429422153366936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,768,0.0061039999127388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,4096,0.010442666709423065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,5120,0.017310221989949543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,2048,0.015964445140626695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,10240,0.030439111921522353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,256,0.004879111217127906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,65536,0.11105333434210883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,5120,0.03192355566554599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,1536,0.009948444035318162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,65536,0.06595200300216675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,7168,0.01685066686736213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,5120,0.012362666428089142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,6144,0.09147910939322577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,12288,0.05947111050287882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,4096,0.023264888260099623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,1024,0.011928888658682505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,768,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,5120,0.016420443852742512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,512,0.015080889066060385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,2048,1.1774302588568795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,12288,1.3698817359076605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,8192,0.03219288918707106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,12288,0.034684442811542086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,512,0.003343111111058129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,128,0.006660444455014334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,768,0.021106666988796655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,10240,0.03045066528850132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,1536,0.04969244533114963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,3072,0.07373155487908258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,16384,0.6162168714735243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,65536,0.04448089003562927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,10240,0.18870933850606283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,128,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,3072,0.01127022256453832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,10240,0.30562933286031085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,51200,0.054423110352622144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,4096,0.0934915542602539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,3072,0.08242311080296834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,5120,0.016677333248986136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,10240,0.17733332845899794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,3072,0.3285751077863905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,10240,0.1273555490705702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,7168,0.023388443721665278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,768,0.04472800095876058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,51200,0.47479910320705837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,768,0.07879110839631823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,128,0.2248586548699273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,128,0.0031644445326593188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,256,0.004688888788223267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,768,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,512,0.016164445214801364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,8192,0.015103111664454142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,16384,0.47980711195203996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,3072,0.00925244473748737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,2560,0.03659822212325202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,10240,0.015599111715952555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,32,0.010778666370444827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,512,0.19603378242916533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,6144,0.018160000443458557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,1024,0.013624000052611033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,512,0.02898400028546651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,1536,0.20015377468532988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,8192,1.0555013020833333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,2048,0.009742221898502773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,8192,0.2772737873925103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,2560,0.016025778320100572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,3584,0.03484444485770331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,2048,0.00721955547730128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,51200,0.010246222217877706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,2048,0.03506133291456435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,51200,1.5500621795654297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,512,0.009672889278994666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,4096,0.02604355580276913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,2048,0.013474666410022311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,7168,0.013329777452680798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,768,0.009017777939637503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,5120,0.023177777727444965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,3072,0.0069448889957533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,4096,0.08065688610076904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,2048,0.07115911112891303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,2048,0.061420447296566434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,16384,0.12387288941277398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,3584,0.07725333505206637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,5120,0.015597333510716757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,1024,0.0041333333485656315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,256,0.0030666666312350165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,8192,0.06739466720157199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,51200,0.3239733378092448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,8192,0.8130302429199219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,5120,0.15810755888621011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,65536,0.024127999941507976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,16384,0.6202630996704102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,10240,0.18001066313849556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,256,0.013284444808959961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,3072,0.008041777544551426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,3584,0.055580443806118436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,1536,0.039854221873813205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,2560,0.05176977647675408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,16384,0.023266666465335425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,12288,0.3488773239983453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,3584,0.013359111216333179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,32,0.13942844337887234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,4096,0.009559111462699043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,64,0.005145777844720417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,10240,0.022406222091780767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,51200,7.1620432535807295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,4096,0.3799866570366754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,12288,0.043122665749655835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,256,0.08641778098212348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,32,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,5120,0.024650666448805068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,3072,0.6009270879957411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,512,0.015751111838552687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,5120,0.45180797576904297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,65536,0.1348924504386054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,10240,0.2038844426472982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,4096,0.19477066728803846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,4096,0.027577777703603108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,65536,12288,2.8779822455512156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,51200,0.40196977721320254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,2048,0.011504888534545898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,7168,0.07191911008622912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,3072,0.010024000373151567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,16384,0.06037511428197225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,128,0.00416088890698221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,4096,0.008904000123341879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,12288,0.022395556171735127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,3072,0.752869341108534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,768,0.005771555420425203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,7168,0.07850399944517347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,51200,0.36233067512512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,1024,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,3584,0.02232355541653103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,51200,1.5219110912746852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,128,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,65536,0.12026311291588677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,256,0.0058319999112023254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,10240,0.016401777664820354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,3584,0.009087111386987898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,5120,0.016730666160583496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,51200,0.4069217840830485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,6144,0.16306043995751276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,64,0.0084915558497111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,2560,0.005976000179847081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,2048,0.012717333104875354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,3072,0.01200977795653873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,1536,0.01089777797460556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,3072,0.05843822161356608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,8192,0.018495111001862418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,128,0.0028373334142896864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,65536,0.07354844278759427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,3584,0.010749333434634738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,12288,0.040864000717798867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,8192,0.011688888900809817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,512,0.015112888481881885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,5120,0.011352889239788055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,128,0.06309600008858575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,12288,0.39152712292141384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,8192,0.006835555450783835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,1536,0.027948445743984644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,10240,0.04670133193333944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,51200,3.6489770677354603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,2560,0.007724444071451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,768,0.007840889195601145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,10240,0.04772888951831394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,3072,0.007224889265166388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,7168,0.0757146676381429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,4096,0.009460444251696268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,3584,0.01050222251150343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,16384,0.009688888986905416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,10240,0.05012889040840996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,2048,0.930538707309299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,6144,0.03061422374513414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,16384,0.6772898038228353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,256,0.03850933247142368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,3584,0.01368977791733212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,65536,9.56239488389757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,10240,0.3422808912065294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,16384,0.02647022240691715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,128,0.01163111130396525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,2560,0.22459199693467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,65536,0.011879111329714457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,256,0.005599110904667113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,256,0.005496888938877318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,512,0.005459555735190709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,3072,0.021379555265108745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,256,0.003576888806290097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,64,0.0027626665929953256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,1024,0.02348800003528595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,6144,0.06264888577991061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,65536,0.3032951090070936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,3072,0.009279111193286048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,12288,0.16352621714274088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,16384,0.40959376758999294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,6144,0.4142613410949707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,128,0.005759999983840519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,32,0.021770666042963665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,1536,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,768,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,32,0.01093777765830358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,65536,2.018550237019857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,4096,0.042840000655916005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,128,0.004851555658711327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,256,0.003608889049953885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,4096,0.013593778014183044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,2560,0.1522595617506239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,8192,0.18863644864824083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,16384,0.03964889049530029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,16384,0.11338488260904948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,256,0.00526577772365676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,10240,0.1020088858074612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,768,0.07880622148513794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,1536,0.10836178064346313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,51200,0.05338044299019707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,1536,0.024118221468395654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,768,0.007672888537247975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,51200,2.1022720336914062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,16384,0.0385155545340644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,2560,0.010672888822025724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,3584,0.026985777748955622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,2048,0.01278222186697854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,12288,0.08138044675191243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,768,0.020810667011472914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,65536,0.08918933073679607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,3584,0.01865244408448537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,6144,0.7630542119344076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,4096,0.008197333249780867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,16384,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,2560,0.025868443979157343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,512,0.04596622122658623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,128,0.004233777936961916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,6144,0.12055199676089817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,2048,0.04621422290802002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,512,0.005950222412745158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,64,0.006785777707894643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,32,0.005167110926575131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,1024,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,5120,0.17122666041056314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,8192,0.26341332329644096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,65536,0.4537955390082465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,1024,0.01646844380431705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,1024,0.011310222248236338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,16384,0.03762400150299072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,5120,0.009887110855844285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,3584,0.012355555262830524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,2048,0.010634666515721215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,4096,0.02382044494152069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,6144,0.013779555757840475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,7168,0.4292613400353326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,51200,0.13837867312961155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,5120,12288,0.21223644415537515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,32,0.006692444284756978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,6144,0.16719467110104033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,10240,0.020962667134073045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,7168,0.01812977757718828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,768,0.028612444798151653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,64,0.026259554757012263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,3584,0.015899555550681222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,8192,0.02866311205757989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,16384,0.45035823186238605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,51200,0.02843022346496582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,768,0.006151999864313338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,3072,0.0284586681260003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,512,0.1390977833006117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,32,0.0034346667428811393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,2048,0.0074346669846110875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,6144,0.04318044583002726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,512,0.1727022197511461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,256,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,512,0.0033555556502607134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,10240,1.3566168679131403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,51200,0.1009262204170227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,256,0.004744889007674323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,3584,0.00777333312564426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,2560,0.009884444375832876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,768,0.02087733315096961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,2048,0.03382666574584113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,3072,0.007987555530336168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,4096,0.01771288944615258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,10240,0.3432933224572076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,768,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,32,0.0029102222373088202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,6144,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,65536,0.34457866350809735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,65536,0.18095110522376168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,10240,0.011437333292431302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,7168,0.19800799422793916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,768,0.006056889063782162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,64,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,32,0.002597333242495855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,10240,0.23753688070509169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,16384,0.02433066732353634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,64,0.0034026667061779234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,1536,0.01975733372900221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,2048,0.09038044346703424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,6144,0.03015822172164917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,1536,0.008640888664457533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,16384,0.0521048903465271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,3072,0.058303111129336886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,16384,0.06795377863778009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,1536,0.11246844132741292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,6144,0.049207998646630176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,10240,0.045401778486039907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,3072,0.05689689185884264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,65536,0.06899555524190266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,16384,0.022329777479171753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,768,0.005678222411208683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,7168,0.006065777606434292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,64,0.0033608888172441055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,16384,0.04673688941531711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,10240,0.017695110705163743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,3072,0.14973333146837023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,5120,0.04571022258864509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,51200,0.4485440254211426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,4096,0.03427022364404466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,256,0.007296889192528195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,128,0.005658666706747479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,2560,0.044512000348832875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,768,0.013160000244776407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,10240,0.027962668074501887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,32,0.00870222184393141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,8192,0.142941329214308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,12288,0.026738666825824316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,256,0.0041777777175108595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,7168,0.020072888996866014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,32,0.0029297777348094513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,256,0.011191111471917896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,7168,0.2644960085550944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,16384,0.157334221733941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,64,0.0026826666047175727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,128,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,3584,0.00889066689544254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,768,0.008075555165608725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,3584,0.1142159965303209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,64,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,512,0.005519111123349931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,10240,0.03995910949177212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,8192,0.06652355856365628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,12288,0.024136889312002394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,64,0.0029075555503368378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,6144,0.11372000641292995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,3584,0.12449599636925592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,12288,0.019334221879641216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,3584,0.04910044537650215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,1024,0.005714666512277391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,51200,2.4917714860704208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,64,0.008856000171767341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,1024,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,3072,0.008629333641793992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,5120,0.011779555843936073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,2048,0.012568000290128918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,6144,0.006711111300521427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,6144,0.015051555302408008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,65536,0.18930578231811523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,1536,0.024686222275098164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,2560,0.01962577799956004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,128,0.008908444808589088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,51200,5.664794498019749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,8192,0.041987554894553296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,12288,0.5759101973639594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,8192,0.016537777251667447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,1024,0.004892444445027245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,3072,0.011353777514563667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,6144,0.02308533257908291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,256,0.005022222383154763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,128,0.02179555594921112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,5120,14.782523261176216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,3584,1.1386950810750325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,1536,0.03528088993496365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,1024,0.005059555586841372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,7168,0.05388622151480781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,10240,0.020450666546821594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,8192,0.03767733441458808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,4096,0.03612533211708069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,7168,0.1569475597805447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,2560,0.3091546694437663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,6144,0.013883555928866068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,768,0.12988800472683376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,3584,0.010943999720944298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,1536,0.006877333339717653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,2560,0.0711137784851922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,128,0.005809777726729711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,768,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,3584,0.2325946754879422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,51200,0.09490578042136298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,4096,0.6117315292358398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,4096,0.09828977452384101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,5120,0.4024222162034776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,3584,0.025144888295067683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,512,0.08532622125413682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,1024,0.024838222397698298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,3584,0.010885333021481832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,12288,0.05979733334647285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,3072,0.03491733471552531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,768,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,3584,0.016139555308553908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,7168,0.012437333663304647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,64,0.032481776343451604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,3584,0.010517333116796283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,512,0.00851733320289188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,7168,0.007575111256705389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,3072,0.026870222555266485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,7168,0.014504000544548035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,6144,0.04909333255555895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,10240,0.2673742241329617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,64,0.0052844443255000645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,2560,0.009025777379671732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,2560,0.021957332889238994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,2048,0.02146044373512268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,128,0.0069217777086628815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,10240,0.07883199718263414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,7168,0.017873777283562552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,256,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,6144,0.19990221659342447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,5120,0.08455999692281087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,4096,0.009011555876996782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,6144,0.013939556148317126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,1024,0.029716445340050593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,128,0.025639999243948195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,10240,0.10342133045196533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,3584,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,128,0.03412177827623155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,2560,0.6386266814337836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,5120,0.04158400164710151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,2560,0.014856000741322836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,1024,0.024854222933451336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,51200,0.27029866642422146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,256,0.02077244387732612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,1536,0.006590222318967183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,128,0.008203555312421586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,3584,0.008173333273993598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,128,0.004401777767472797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,16384,2.20560540093316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,12288,1.1812426249186199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,8192,0.33981598748101133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,1536,0.8951395882500542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,5120,0.009959110783206092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,5120,0.685364458296034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,10240,0.02296977738539378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,5120,0.013328889177905189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,10240,0.019521777828534443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,65536,3.0482364230685763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,2560,0.6275555292765299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,5120,0.018069333500332303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,256,0.006161777923504512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,3584,0.012190221912331052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,3072,0.030589332183202107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,7168,0.021094222863515217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,12288,0.02648799949222141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,128,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,16384,0.43450578053792316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,512,0.01162755572133594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,64,0.006154666758245892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,5120,0.034879999028311834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,1024,0.015896888242827523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,51200,0.4918035401238336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,16384,0.4418337874942356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,1024,0.011531555818186866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,6144,0.04362577862209744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,7168,0.9815307193332248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,7168,0.024872000018755596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,16384,0.050717333952585854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,128,0.25338133176167804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,128,0.003678222083383136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,10240,0.21919377644856772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,5120,0.01038755559258991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,768,0.019173334042231243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,10240,0.44852887259589297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,16384,0.07657866345511542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,512,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,12288,0.3718986776140001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,5120,1.6019875208536785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,4096,0.013950222068362765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,8192,0.007877333296669854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,8192,0.1499484380086263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,512,0.00794222205877304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,768,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,2560,0.3406533400217692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,10240,0.1875120004018148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,1536,0.011237333218256632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,65536,1.1699928707546658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,3072,0.012140444583363004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,128,0.005930666708283954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,64,0.004986666556861666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,512,0.020640000700950623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,5120,0.855314678615994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,1024,0.10196444723341201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,1024,0.0064399999876817065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,2560,0.07484000258975558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,5120,0.0222044438123703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,6144,0.011872000164455838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,8192,0.020620443754725985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,16384,0.2410257657368978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,3072,0.12714755535125732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,12288,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,1536,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,5120,0.03787644373046027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,12288,0.1605964501698812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,4096,0.6858462227715386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,51200,0.19301954905192056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,768,0.006664000037643645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,7168,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,1024,0.3663635518815782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,128,0.00425866660144594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,3072,0.04496533340877957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,2560,0.013925333817799887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,1536,0.4534293280707465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,512,0.0796239972114563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,64,0.0025439999169773525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,1536,0.018538667096032035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,2560,0.005948444621430502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,8192,0.20680177211761475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,2560,0.058522668149736196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,4096,0.38449690077039933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,5120,0.013367999758985309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,7168,0.3437679873572455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,4096,0.0626115534040663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,10240,0.015678221980730694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,1024,0.0277839998404185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,4096,0.05481600099139743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,3072,0.0351893338892195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,3072,0.0110880004035102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,32,0.06274577644136217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,32,51200,0.013256889250543384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,64,1536,0.01460444430510203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,51200,1.942411634657118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,768,0.024554666545655992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,3072,0.044381333722008594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,7168,0.021781333618693884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,8192,0.06215555800331963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,4096,0.011412444214026133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,5120,0.13768800099690756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,128,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,3072,0.007836444510353936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,12288,0.12011288272009955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,10240,0.03228622343805101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,5120,0.20279288291931152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,32,0.003121777748068174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,64,0.008550222549173567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,5120,0.03929155402713352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,16384,0.041486223538716636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,5120,2.101620356241862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,8192,0.01833066675398085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,6144,0.00945955514907837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,1024,0.04190133346451653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,32,1.950842751397027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,3072,0.030476444297366675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,3584,0.04746577805942959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,12288,0.06367644336488512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,5120,0.013842666314707862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,256,0.01736177835199568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,3584,0.013251555462678274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,10240,0.04047200083732605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,128,0.005166222237878376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,4096,0.01645511057641771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,8192,0.17248888810475668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,768,0.024259555670950148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,3584,0.0756675534778171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,512,0.00835111074977451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,1536,0.05741155809826321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,5120,0.06644799974229601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,7168,0.1857768827014499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,12288,0.006108444597986009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,4096,0.01402311192618476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,2048,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,256,0.0027955555253558685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,3584,7.190397474500869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,64,0.00794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,51200,1.4763831032647028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,32,0.008434666527642144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,1536,0.02453244394726223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,768,0.021839110387696162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,10240,0.09453777472178142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,256,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,2560,0.07856977648205228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,256,0.026583110292752583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,8192,0.018796443939208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,7168,0.028365333875020344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,768,0.02576711111598545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,5120,0.11033066776063706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,7168,0.014984000060293408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,65536,0.5350480079650879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,4096,0.015973332855436537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,32,0.004197333421972063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,8192,0.02094399929046631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,256,0.03242044316397773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,768,0.061308443546295166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,5120,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,5120,0.11212710539499919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,5120,0.0244159996509552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,6144,0.013360888593726687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,7168,0.030340443054835003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,3072,0.03032711148262024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,3072,0.009935110807418823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,3072,0.06239644686381022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,16384,0.05681866407394409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,2560,0.13017955091264513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,1536,0.2808489004770915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,768,0.006292444550328785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,128,0.004709333181381226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,1024,0.01033688916100396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,128,0.0058746664888328975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,512,0.016805332567956712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,3072,0.009516444471147325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,2560,0.015286222100257874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,2048,0.05532088875770569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,1024,0.016516443755891588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,32,0.004222222086456087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,8192,0.02353511088424259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,12288,0.02648711038960351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,7168,0.011052444577217102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,12288,1.1413218180338542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,7168,0.060384889443715416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,4096,0.022234666678640578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,5120,0.045144889089796275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,64,0.007428444094128079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,2048,0.018999111321237352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,5120,0.026584888497988384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,10240,0.07356355587641399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,768,0.005868444426192178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,32,0.02687377731005351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,2560,1.9431076049804688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,5120,0.0341306659910414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,51200,3.062391916910807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,4096,0.00999111102686988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,16384,0.05115288827154371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,1536,0.0673475530412462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,256,0.019104000594880845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,2560,0.2329937881893582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,4096,0.012551111479600271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,51200,0.7320639822218152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,65536,1.3819519678751628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,7168,0.02409422232045068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,51200,0.28905243343777126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,2048,0.07947644260194567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,12288,256,0.007199111084143321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,7168,0.02157955533928341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,2048,0.13349689377678767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,6144,0.47795820236206055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,12288,0.051177776522106595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,65536,0.17812800407409668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,1024,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,4096,0.006861333217885759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,32,0.005065777649482091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,5120,0.895567152235243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,7168,3.1179652743869357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,6144,0.11303199662102593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,16384,0.024150222539901733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,3584,0.009166222479608323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,1024,0.03659910957018534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,512,0.015272888872358533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,16384,0.13958311080932617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,768,0.006603555546866522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,128,0.007499555746714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,6144,0.2210222217771742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,16384,0.03760800096723769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,12288,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,65536,0.009870222045315636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,256,0.0032826666202810076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,768,0.014688889185587565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,4096,0.11416799492306179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,8192,0.6331902080112034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,2560,0.07200088765886095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,5120,0.05407999952634176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,51200,0.4839946428934733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,768,0.004052444464630551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,16384,0.4527769088745117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,65536,0.04594666759173075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,51200,0.06741778055826823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,12288,0.03131555517514547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,65536,0.07705333497789171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,8192,0.013325333595275879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,7168,0.039078222380744085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,7168,0.010989333192507425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,768,0.011720889144473605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,7168,1.387695100572374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,1024,0.016466667254765827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,2048,0.0845866666899787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,16384,0.025441777375009324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,32,0.00927111092540953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,7168,0.1299066676033868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,4096,0.011339555184046427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,2048,0.006950221955776215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,7168,0.05834666887919108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,1536,0.025693333811230127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,2048,0.005579555614127054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,16384,0.06884355677498712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,3072,0.0127697777416971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,1024,0.2609013451470269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,65536,0.06746221913231744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,10240,0.41130489773220485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,256,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,3072,0.08746755785412258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,65536,4.2759450276692705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,4096,0.010815111299355825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,32,0.006538666784763336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,65536,0.10732444127400716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,16384,0.052863111098607383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,1536,0.07028266456392077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,768,0.013453333742088742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,16384,0.3279946645100911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,6144,0.006536888993448681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,10240,0.03789155681927999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,3072,0.16787644227345785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,128,0.004038222134113312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,5120,12288,0.02611200014750163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,3072,0.006462222172154321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,16384,0.24913867314656576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,256,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,6144,0.011600000162919363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,51200,0.8589590920342339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,3072,0.18600800302293566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,2560,0.016885333591037326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,128,0.0036799998746977914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,7168,0.04829511046409607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,128,0.006793777975771163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,5120,0.020851555797788832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,10240,0.1443155606587728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,768,0.005471111171775394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,4096,0.01756088932355245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,512,0.018040888839297824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,5120,0.04243200023969015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,8192,0.014353776971499125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,3072,0.008367111285527548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,1536,0.01354311075475481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,12288,0.1078195571899414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,64,0.0032631111227803757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,8192,0.016593777471118502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,5120,0.7113128768073188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,4096,0.06599289178848267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,2560,0.26453155941433376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,2048,0.013927999469969006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,10240,0.04069777660899692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,5120,0.01033333357837465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,32,0.004108444270160464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,2048,0.012884444660610624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,6144,0.06304266717698839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,6144,0.19715732998318145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,8192,0.04862577716509501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,1536,0.08447111315197414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,7168,0.10334399673673843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,4096,0.02646044393380483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,5120,0.0676142242219713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,2048,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,2560,0.011596444580290051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,10240,0.14672356181674534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,64,0.42463111877441406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,65536,3.8391058180067272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,128,0.009464888936943477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,12288,0.03584444522857666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,4096,0.014322666658295525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,6144,0.4463689062330458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,512,0.006108444597986009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,12288,0.022966222630606756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,256,0.003759110967318217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,6144,0.015207999282413058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,3584,0.04807200034459432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,3072,0.011427555647161273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,12288,0.08451110786861843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,12288,0.18518577681647408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,128,0.00790666706032223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,16384,0.11255378193325466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,2048,0.007704888780911763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,768,0.003961777935425441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,8192,0.01983644399378035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,768,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,128,0.006629333313968446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,128,0.005899555567238066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,2560,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,2560,0.013007111019558378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,65536,1.0305840174357097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,65536,0.47154664993286133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,1024,0.007771555748250749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,2048,0.008796444369686974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,8192,0.06226578023698595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,16384,0.021158221695158217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,128,0.009696000152164036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,128,0.004943111290534337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,51200,1.6487378014458551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,768,0.026009778181711834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,3072,0.007633777956167857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,4096,0.05968533621893989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,256,0.04670133193333944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,5120,0.022841778066423204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,6144,0.03661244445376926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,1024,0.008920888933870528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,4096,0.009344889058007134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,32,0.017673777209387887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,4096,0.031197332673602637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,1536,0.06568444437450834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,5120,0.006783111227883234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,2560,0.006805333412355847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,64,0.0073315559162033936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,768,0.1531777779261271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,256,0.016155555844306946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,51200,0.058315555254618325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,3072,0.0691831111907959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,2560,0.09862666659884983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,16384,0.31908265749613446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,1024,0.014264000786675347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,10240,0.41051022211710614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,10240,0.01847999956872728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,1024,0.006528888725572162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,64,0.002983999955985281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,51200,0.14929955535464817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,3072,0.007965332931942409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,10240,0.010840888652536603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,2560,0.03302844365437826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,128,0.0027964444210131965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,2560,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,7168,0.6016826629638672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,6144,0.021384000778198242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,16384,0.40013689464992946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,7168,0.2026346657011244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,2048,0.1952577829360962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,1536,0.00996888925631841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,128,0.008340444829728868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,51200,1.365028487311469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,512,0.006823110911581252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,1536,0.01574133336544037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,128,0.0036675557494163513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,6144,0.007712889048788283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,8192,0.02601066728432973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,768,0.006734222173690796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,8192,0.040249778164757624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,3072,0.12954755624135336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,3072,0.011015999648306104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,65536,0.09871021906534831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,65536,10.066018846299913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,51200,0.5911359786987305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,32,0.003673777812057071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,256,0.006080889039569431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,256,0.0041831110914548235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,10240,0.02653955585426754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,64,0.007139555282062954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,5120,0.08570400211546157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,1024,0.01143022212717268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,3072,0.04832266767819723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,768,0.023192000057962205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,5120,0.08781333102120294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,12288,0.2372586727142334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,6144,0.051242666112052075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,51200,0.030850665436850652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,32,0.002487999904486868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,65536,0.7528737915886773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,8192,0.41523467169867617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,4096,0.14770132965511745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,65536,0.21335555447472465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,6144,0.03364888827006022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,512,0.007061333292060428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,256,0.006847111301289664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,128,0.013208889298968844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,512,0.021926222576035395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,1024,0.009855999714798397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,512,0.042591998974482216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,8192,0.12447644604576959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,3072,0.20449154906802705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,5120,0.01567555632856157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,65536,0.014491554763582019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,51200,3.379343032836914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,1536,0.1272853347990248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,2560,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,4096,1536,0.08427288797166611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,3584,0.019567999574873183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,3584,0.034726222356160484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,7168,0.006628444625271692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,65536,0.16721777121225992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,512,3072,0.012149333126015134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,1024,0.014947555131382413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,10240,0.01832177738348643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,1536,0.021127111381954614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,512,0.01055466632048289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,128,0.3947448995378282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,7168,0.06045599778493246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,128,0.0042071110672420925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,8192,0.012267555627557965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,4096,0.08883288833830093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,2560,0.007027555671003129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,3584,0.677236451043023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,768,0.03959822323587205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,5120,0.0076142218377855085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,7168,0.02504088812404209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,128,0.031110223796632554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,768,0.006816000160243776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,10240,0.02437155610985226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,12288,0.0845057765642802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,768,0.019700444406933255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,2560,0.007999111380841997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,8192,0.019998222589492798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,7168,0.01607911123169793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,12288,0.0092275556590822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,8192,0.029784000582165186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,2560,0.013629333012633853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,8192,2.856704923841688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,7168,1.9668045043945312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,128,0.009531555904282464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,512,0.004550222307443619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,12288,0.32029954592386883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,4096,0.00979644474056032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,5120,0.23196445571051705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,512,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,12288,0.02492622203297085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,256,0.004591111093759537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,2048,0.02561155623859829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,2048,0.1214560005399916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,1536,0.01684799955950843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,6144,0.552436457739936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,51200,1.5843138164944117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,2048,0.007854222423500484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,2048,0.025589333640204534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,1536,0.049485332436031766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,1536,0.011137777732478248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,512,0.005680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,1536,0.017867555220921833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,7168,0.04307644565900167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,65536,0.07202311356862386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,128,0.007333333293596904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,128,0.026884444885783728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,51200,3.3513832092285156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,5120,0.0914906660715739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,512,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,65536,0.48543288972642684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,51200,1.507615089416504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,10240,0.020656888683636982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,1024,0.007845333053006066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,10240,0.009723555710580613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,2560,0.10148533185323079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,5120,0.01769244505299462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,16384,0.33160000377231175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,7168,0.17000444730122885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,10240,0.018481777773963082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,3584,0.015033778217103748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,8192,0.021624000536070928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,128,0.009352888498041365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,51200,48.29393174913195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,7168,6144,2.7118409474690757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,65536,1.0222053527832031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,2048,0.012688000169065265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,2560,1.0165964762369792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,768,0.0036311112344264984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,256,0.00489777781897121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,10240,0.010991999672518836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,768,0.028504000769721136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,5120,0.2633333206176758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,32,0.006464889066086874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,2560,0.045759111642837524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,512,0.041638222005632185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,128,0.11735110812717015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,32,0.008629333641793992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,4096,0.010544000400437249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,12288,0.09991288847393459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,768,0.00609688874748018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,10240,0.4670115576850043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,8192,0.11710222562154134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,10240,0.08405955632527669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,10240,0.03319022059440613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,1536,0.04192177785767449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,6144,0.02180800007449256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,3584,0.03456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,512,0.004914666629499859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,7168,0.020603555772039626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,7168,0.5291475719875759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,65536,0.15022755993737116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,51200,3.785071055094401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,16384,0.8251742257012261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,10240,2.5576400756835938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,6144,0.08010044362809923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,4096,0.05682310793134901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,5120,0.5548035303751627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,65536,0.10933155483669704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,51200,2.1785235934787326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,32,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,12288,3.7842665778266062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,4096,0.030679110023710463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,4096,0.021342222889264423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,2048,0.03282489048110114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,12288,0.007143110864692264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,8192,0.08153777652316622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,4096,0.013337777720557319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,8192,0.49332088894314235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,65536,0.19570044676462808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,256,0.0036951113078329298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,128,0.007037333316273159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,768,128,0.004263111286693149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,512,0.009272889130645329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,256,0.004480888860093223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,6144,0.03367111086845398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,256,0.01088177743885252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,6144,0.034303112162484065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,768,0.2686186631520589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,5120,0.10834755500157674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,128,0.0029306666304667792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,2560,0.07859022087521024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,1536,0.029157333903842505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,2560,0.011786667009194693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,256,0.00562755556570159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,7168,0.018431110514534842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,16384,0.1424835522969564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,8192,0.014422222971916199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,3584,0.05699555741416084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,51200,0.6668968730502658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,5120,0.9395457373725044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,65536,0.1720942258834839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,6144,0.028757333755493164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,10240,0.34080712000528973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,3584,0.011410666836632622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,64,0.01183199965291553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,128,0.0027626665929953256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,5120,0.041823112302356295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,3584,0.009977777798970541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,7168,0.013568888935777875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,128,0.0063279999627007385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,128,0.009752888646390703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,2560,0.030829333596759375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,64,0.030551999807357788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,6144,0.01092888911565145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,16384,7168,0.03837155633502536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,3072,0.011217777927716574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,32,0.019128888845443726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,32,0.008234666453467475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,128,0.018498667412334018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,1024,0.012061333490742577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,4096,0.6442524592081705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,32,0.0034471110751231513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,8192,0.013633777697881063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,7168,1.8703422546386719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,51200,0.11552622583177354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,3072,0.018186666899257235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,51200,34.69521755642361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,6144,0.030392888519499037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,128,0.006083555519580841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,256,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,3584,0.03693688909212748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,1536,0.15174489551120335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,12288,0.044234666559431285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,1536,0.05829066700405545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,2048,0.0748986668056912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,3072,0.01053066634469562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,768,0.005679111099905438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,1024,0.00795377790927887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,6144,0.02719555629624261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,128,0.007861333588759104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,512,0.009211555951171452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,2560,0.01934400035275353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,5120,0.05920444594489204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,6144,0.02851288848453098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,512,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,3072,0.006940444310506185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,32,0.16711733076307508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,5120,0.018618666463428073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,5120,3.4757031334771047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,16384,2.2313573625352645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,65536,0.1942666636572944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,16384,0.0444124440352122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,1024,0.00871377769443724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,64,0.02964622113439772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,3584,0.010257778068383535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,256,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,65536,0.14224444495307073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,3584,0.0740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,6144,0.09383200274573432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,51200,0.2310240003797743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,5120,0.14095732900831434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,65536,0.14872445000542536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,2560,0.09144622087478638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,1536,0.01531022290388743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,128,0.03289688958062066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,7168,0.05068622363938225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,16384,0.08695377906163533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,7168,0.06297689014010958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,768,0.009213333328564962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,16384,0.14973510636223689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,1024,0.031707555055618286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,6144,0.09290755457348293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,64,0.00360622215602133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,2560,0.0565155545870463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,7168,0.35749067200554746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,2560,0.020042666130595736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,2560,0.10238222281138103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,7168,0.03365600109100342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,12288,0.05871555540296766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,3584,0.4728719923231337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,10240,0.026491555902693007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,3072,0.7677404615614148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,51200,0.1438328954908583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,12288,0.16993244489034018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,128,0.005480888817045424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,51200,0.11997689141167535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,4096,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,7168,0.021612443857722815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,10240,0.01953066719902886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,256,0.1829528941048516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,51200,0.11662756072150336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,256,0.005136888888147143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,51200,0.5150871276855469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,2048,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,768,0.007815111014578078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,128,0.0068764446510208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,12288,0.010855110983053843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,128,0.01960088809331258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,51200,0.09693333175447251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,16384,0.16085422039031982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,64,0.019508444600635104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,7168,0.038951999611324735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,1024,0.015493333339691162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,6144,0.03310755557484097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,10240,2.227420383029514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,1536,0.7490062183803983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,2560,0.08483644326527913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,10240,1.2516933017306857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,3072,0.02199911077817281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,1024,0.01091644416252772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,3072,0.2815973228878445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,6144,0.006258666515350342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,5120,0.011968889170222811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,7168,0.03315022256639268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,16384,0.2516000005933973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,2048,0.011680888632933298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,64,0.0028888889484935333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,8192,0.4140560097164578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,8192,0.22639645470513237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,2560,0.009434666898515489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,51200,0.6357475386725532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,5120,768,0.006239111224810283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,32,0.0027831111931138565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,128,0.019724445210562814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,6144,0.018221333622932434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,8192,0.24142577913072374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,4096,0.18824356132083467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,51200,0.7047368685404459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,10240,0.0333280000421736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,128,0.00666222224632899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,768,0.010136889086829292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,128,0.00572533326016532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,2560,0.016926222377353244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,3072,0.04189244574970669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,65536,0.07394222418467204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,3072,0.015391111373901367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,10240,0.07770311170154147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,12288,0.6029120021396214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,8192,0.03150044547186957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,3584,0.011927111281288994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,5120,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,12288,0.01812799937195248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,256,0.013946667313575745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,6144,0.032820443312327065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,3584,0.012074666718641916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,2048,0.00701155596309238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,10240,0.04937600096066793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,128,0.0032142222755485107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,3584,0.07740444607204862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,12288,0.06940355565812853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,3584,0.04672799838913811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,64,0.08291288879182604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,768,0.02130844526820713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,3072,0.011658666862381829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,768,0.6021768781873915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,512,0.006141333530346553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,1536,0.007172444628344641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,512,0.00702400008837382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,128,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,768,0.00812711069981257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,51200,1.3583279715643988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,12288,0.771866692437066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,3072,0.1084666649500529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,3584,0.07138044304317899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,4096,0.06869422064887153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,10240,0.0450853341155582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,1536,0.008014221986134848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,512,0.02422933280467987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,1536,0.504691547817654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,2048,0.1521137820349799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,12288,0.05346222387419807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,3584,0.020633776982625324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,6144,0.06046666701634725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,1024,0.004355555607212914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,8192,0.0074924445814556545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,4096,0.014244443840450711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2048,256,0.004573333180612988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,3072,0.007830222447713217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,1024,0.005995555470387141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,8192,0.015820443630218506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,7168,0.28091200192769367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,256,0.003056888985964987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,128,0.0027919999427265594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,3072,0.011679111255539788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,768,0.010272888673676385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,3072,0.015662221444977652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,128,0.05474933319621616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,256,0.0033191111352708605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,12288,2.54951900906033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,10240,0.3100586732228597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,256,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,2048,0.05997955799102783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,7168,0.39186755816141766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,7168,0.020115555988417733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,7168,0.04357777701483833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,6144,0.20509066846635607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,8192,0.010235555469989777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,16384,0.8666728867424859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,32,0.003560888684458203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,2048,0.008782222039169734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,3584,0.016452444924248587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,256,0.008293333152929941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,32,0.0074746666683091064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,32,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,8192,2.496172375149197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,3072,0.016957332690556843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,8192,0.006665777828958299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,3584,0.011110222174061669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,256,0.05797244442833794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,1024,0.017367111312018502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,5120,0.07662577761544122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,5120,0.06299110915925768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,7168,0.08660888671875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,3072,0.01370844410525428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,16384,0.021520000365045335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,128,0.006446222050322427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,1536,0.0064488889442549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,4096,0.016993777619467843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,65536,13.673855251736112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,10240,0.031927112076017596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,3584,0.027638221780459087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,51200,0.2719448937310113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,6144,2.8483395046657987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,768,0.055330667230818004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,3584,0.10301688644621108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,10240,0.5325653288099501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,128,0.018222222725550335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,64,0.003026666740576426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,8192,0.07143644491831462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,256,0.057308442062801786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,2560,0.08268177509307861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,2048,0.17345421844058564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,12288,0.020085333122147452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,64,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,64,0.0028328889360030494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,4096,0.03128977616628011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,8192,0.021631111701329548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,1536,0.21922043959299722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,1536,0.718140443166097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,6144,0.010229333407349056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,2560,1.0465075174967449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,10240,0.02167822255028619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,1024,0.04090933336151971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,2048,0.21320888731214735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,5120,0.021550221575631037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,64,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,3072,0.006711999989218182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,1536,0.008780444661776224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,51200,2.0623395707872176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,1024,0.033173332611719765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,1024,0.022964444425370958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,8192,0.018443556295500863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,1536,0.09303200244903564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,64,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,1536,0.016354666815863717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,3072,0.01069866700304879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,10240,0.015642666154437594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,128,0.005450666778617435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,2560,0.06103377872043186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,1536,0.0058400001790788435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,2560,0.8263955646091037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,65536,0.12833333015441895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,64,0.003035555490189128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,2560,0.2384293344285753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,768,0.06694222158855863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,128,0.005006222261322869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,512,1024,0.0052782222628593445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,65536,0.1649964385562473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,6144,0.9296418295966254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,2560,0.011367999845080905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,128,0.0037191112836201987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,65536,3.6201655069986978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,7168,0.28536534309387207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,65536,0.019779556327395968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,768,0.007851555943489075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,3584,0.010103111465771994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,16384,0.31870222091674805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,1536,0.011072888970375061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,10240,0.06095911396874321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,2560,0.00675999994079272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,8192,0.040735999743143715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,768,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,8192,0.020432889461517334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,51200,0.7231689029269748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,6144,0.1209146711561415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,3072,0.01200177768866221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,2048,0.025560888979170058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,64,0.006970666348934174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,2560,0.019919999771647982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,2048,0.03443288803100586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,5120,0.07716977596282959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,3584,0.2736577722761366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,768,0.007729777859316931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,512,0.005199111170238919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,32,0.002567111204067866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,4096,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,5120,0.010463111102581024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,1536,0.00646311127477222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,1024,0.004652444273233414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,65536,0.36397422684563535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,1536,0.006881777611043718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,256,0.024030221833123103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,16384,0.022848000129063923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,7168,0.049695998430252075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,65536,4.409210628933377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,10240,0.23610756132337782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,768,0.07167288992140028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,8192,0.006513777706358168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,12288,0.9186124801635742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,512,0.00810933361450831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,32,0.0031022222505675424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,6144,2.4780169592963324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,64,0.00886933339966668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,128,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,768,0.0331582228342692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,1536,0.6099937756856283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,3072,0.019555555449591745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,3584,0.011401777466138205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,1536,0.21240444978078207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,128,0.02149244480662876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,3584,0.014649778604507446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,1536,0.019139554765489366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,64,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,4096,0.03683022326893277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,64,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,256,0.02420444455411699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,65536,3.0152276357014975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,3584,0.013257777525318993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,128,0.006671999891599019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,4096,0.02944533361328973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,3072,1.4479493035210504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,4096,0.02347911066479153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,2560,0.02683288852373759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,1024,0.014825777875052558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,6144,0.013088000317414602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,65536,3.0151324801974826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,8192,0.03410311208830939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,256,0.0379688905345069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,1536,0.28655465443929035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,3584,0.04978488882382711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,7168,0.32500889566209584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,65536,1.576708475748698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,5120,0.015950222810109455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,1536,0.009827555881606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,3584,0.009240888886981541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,12288,0.01830755505296919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,8192,0.1085253357887268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,768,0.04668622215588888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,2560,0.029924445682101782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,128,0.005653333332803514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,6144,0.02697511182890998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,512,0.010819555156760745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,256,0.009224889179070791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,256,0.005098666581842634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,1024,0.010384000009960597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,7168,0.012096000214417776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,8192,0.02214755614598592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,1024,0.005119110975000593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,51200,0.05128622386190626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,2560,0.6052906778123643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,256,0.00648266656531228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,32,0.01092888911565145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,3584,1.4726720386081273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,2048,0.010598222414652506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,2560,0.019297776950730216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,768,0.004898666507667965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,51200,0.8039555549621582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,256,0.2286720010969374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,5120,0.07000977463192411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,2048,0.006813333266311222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,32,0.005436444448100195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,768,0.00553688903649648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,10240,0.6288248697916666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,3584,0.2304826577504476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,6144,0.025164445241292317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,1024,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,512,0.19320889314015707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,6144,0.050344887706968516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,1536,0.05410933494567871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,128,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,2560,0.01904355486234029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,3072,0.013068444199032254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,8192,0.06488711304134794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,10240,0.05992088715235392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,12288,0.020062221421135795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,51200,0.16047822104560003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,128,0.004495999879307217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,32,0.0029084444459941653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,256,0.003216888962520493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,1536,1.2349209255642362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,5120,0.0699217783080207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2560,51200,0.25218844413757324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,512,0.006588444527652528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,16384,0.03872444563441806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,6144,0.013217777841620974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,4096,0.18618222077687582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,10240,0.5623013178507487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,1024,0.3824160099029541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,65536,0.2935822274949816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,64,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,51200,0.10617599884668986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,16384,0.12375822332170273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,64,0.003101333354910215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,768,0.0064106666379504735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,10240,0.01761244402991401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,51200,10.29864501953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,2560,0.2060257726245456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,4096,0.02068533334467146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,3072,0.178929779264662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,6144,0.6534746487935384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,768,0.007683555285135905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,128,0.009648888475365108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,64,0.020787555310461257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,3584,0.016382222374280293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,10240,0.18221244547102186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,128,0.003633777714437909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,6144,0.010093332992659675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,2560,0.013481777575280933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,128,0.007664889097213745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,2560,0.011184888581434885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,5120,0.22068444887797037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,1536,0.010469333165221743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,256,0.004734222259786394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,32,0.006638222270541721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,128,0.004915555732117759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,128,0.012423111332787408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,65536,1.0534124374389648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,5120,0.35320088598463273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,2048,0.011679111255539788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,128,0.004584889031118817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,256,0.04347911145952013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,4096,0.027315555347336665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,4096,0.02896889050801595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,256,0.006215111249023014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,6144,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,768,0.004573333180612988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,65536,0.08955466747283936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,64,0.0026906666656335196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,10240,0.029016888803905908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,4096,0.274616003036499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,2560,0.12233155303531223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,128,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,65536,0.11077244414223565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,768,0.011492444409264458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,8192,0.15596889124976263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,51200,0.20576443937089708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,64,0.004991111242108875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,128,0.020435555113686454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,3584,0.011507555842399597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,3072,0.017077333397335477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,32,0.006270222365856171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,6144,0.05201066533724467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,512,0.03209599852561951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,512,0.03591288791762458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,64,0.14939822090996638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1024,2048,0.007155555817815993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,7168,0.011005333728260465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,12288,0.2250462108188205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,128,0.0027884443600972495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,5120,0.03567377726236979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,1536,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,5120,8192,0.028597331709331934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,512,0.014952888091405233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,3584,0.12978043821122912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,8192,0.03819377885924445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,3072,0.12364621957143147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,5120,0.16887111133999297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,256,0.0032426667296224167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,12288,0.5390515327453613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,512,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,4096,0.035875555541780256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,5120,0.01328000012371275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,1024,0.004843555390834808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,12288,0.4302195443047418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,3584,0.10467733277214898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,2048,0.014967999524540372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,768,0.011696000066068439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,65536,0.11959821648067898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,65536,0.3324791060553657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,1536,0.04962844318813748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,16384,0.03608444333076477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,8192,0.006795555353164673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,5120,0.009545777406957414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,5120,1.3041999604966905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,256,0.005192000004980299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,51200,0.06957688596513537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3584,128,0.015124445160230001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,512,0.05430488785107931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,3584,0.020901333954599168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,3072,0.11523111661275227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,256,0.03151999910672506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,1024,0.011420444481902652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,1024,0.037385778294669256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,51200,0.04876177840762668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,256,0.01221333361334271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,51200,0.05824266539679634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,32,0.006582222051090664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,7168,0.32090844048394096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,10240,0.06493244568506877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,128,0.0027359999302360746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,3584,0.11656888326009114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,2560,0.023031110564867657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,2560,0.009907555249002244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,256,0.003031999907559819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,512,0.042576889197031655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,512,0.005736888696750005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,3584,0.012038222617573209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,128,0.07723289065890841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,1536,0.004650666481918759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,8192,0.15654755963219538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,2560,0.009493333597977957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,1536,0.06204622321658664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,128,0.015007999208238391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,4096,0.05582133266660902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,5120,0.04673777686225044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,128,0.004507555729813046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,8192,0.11349333657158746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,10240,0.04610222246911791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,6144,0.028188443846172754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,16384,0.42125688658820254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,65536,0.17749599615732828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,4096,0.05657955673005846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,128,12288,0.037913779417673744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,256,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,32,0.0037271111375755737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,768,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,3584,0.18623110983106825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,64,0.007129777636792924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,12288,0.417343987358941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,4096,0.09506577915615505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,64,0.0025537777692079544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,6144,0.013742222554153867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,1536,0.046952890025244824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,2560,0.05057155423694187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,65536,0.2124159998363919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,65536,0.20564888583289254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,7168,0.018457776970333524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,2560,0.036308444208569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,2560,0.09966578086217244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,6144,0.015896000795894198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,256,0.012629333469602795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,2560,0.007500444849332173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,256,0.004495999879307217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,8192,0.026506667335828144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,7168,2.3526168399386935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,768,0.008791111409664154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,512,0.011385777758227454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,3584,0.34509510464138454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,3072,0.0169831116994222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,1536,0.014017777310477363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,256,0.006534222099516127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,3072,0.012348444097571902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,512,0.03737599982155694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,512,0.005320888840489917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,128,0.0874062246746487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,2048,0.045174221197764076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,32,0.0217866665787167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,512,0.43582667244805234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,1024,0.009642666412724389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,5120,0.02975377771589491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,256,0.01145511120557785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,10240,0.03829955392413669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,5120,0.051070223251978554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,7168,0.015056000815497505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,64,0.002975111206372579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,16384,0.543865786658393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,16384,0.1470586723751492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,2048,0.09006577730178833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,12288,4.330694410536024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,5120,0.024413333998786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,6144,0.07488977909088135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,65536,25.262281629774307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,8192,0.24108266830444336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,8192,0.011277333729796939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,2048,0.014559111661381192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,128,0.003096888876623578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,2048,0.08178489075766669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,12288,0.23514755566914877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,3584,0.009312888814343346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,10240,0.008199111455016667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,2048,0.05681600173314413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,128,0.012028444144460889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,256,0.003069333318206999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,2560,0.022982221510675218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,5120,0.05597511265012953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,3072,0.01568800045384301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,65536,0.43373600641886395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,3072,0.01187999960449007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,3584,0.02065333392884996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,512,0.00342755557762252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,7168,0.05741066402859158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,4096,0.011848889291286469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,3072,0.018377777602937486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,2560,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,65536,0.07638311386108398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,65536,0.2325662242041694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,1024,0.008801778157552084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,7168,0.21350133419036865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,3072,0.007902222375075022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,7168,0.12662666373782688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,128,0.006792000184456508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,64,0.005041777673694822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,2048,0.10714222325219049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,5120,0.015967110792795818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,32,0.011767110890812345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,51200,0.08559377988179524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,1024,0.2822035683525933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,1536,0.005626666463083691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,51200,0.22562045521206328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,51200,0.13926755057440862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,8192,0.26578577359517414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,1536,0.013669333524174161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,256,0.11766221788194443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,2048,0.03219111098183526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,4096,0.02276533345381419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,10240,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,1536,0.07190222210354276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,3584,0.21799111366271973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,12288,0.07097955544789632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,3584,0.027062222361564636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,8192,2.2131235334608292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,10240,0.02987288766437107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,1536,0.006526222245560751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,10240,0.9177012973361545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,256,0.004106666478845808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,12288,7.997078789605035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,3584,0.043487111727396645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,1024,0.09186133411195542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,3072,0.015172445111804538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,32,0.0032720000793536506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,1024,0.005889777921968036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,6144,0.2785724533928765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,1536,0.15179733435312906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,8192,0.017394666870435078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,3072,0.007186666958861881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,6144,0.016532444291644625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,8192,0.2135582235124376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,512,0.3631440003712972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,256,0.0032311110860771606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,3072,0.14758222632937962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,8192,4.0066142612033415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,1024,0.006203555398517185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,65536,0.40008089277479386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,256,0.013806222213639153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,8192,4.718006981743707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,8192,0.022210665875011023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,16384,1.07695378197564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,12288,0.23646844757927787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,512,0.004035555654101902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,4096,0.018817777434984844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,16384,0.04840355449252658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,3584,0.012372444073359171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,768,768,0.005798222290145026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,5120,0.13557867209116617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,10240,0.07058844301435682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,128,0.01033866653839747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,3072,0.03487644592920939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,7168,0.01556711064444648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,4096,0.012808889150619507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,10240,0.01186044431395001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,128,0.012281777958075205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,6144,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,512,0.004566222015354368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,4096,0.02001333402262794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,512,0.010528888967302112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,8192,0.1515600019031101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,5120,0.048355556196636625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,512,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,10240,0.259573327170478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,5120,0.006675555474228329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,7168,0.16715733210245767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,768,0.0036675557494163513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,512,0.008564444051848518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,51200,7.6685536702473955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,16384,0.02769600020514594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,10240,0.15497422218322754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,32,0.045216000742382474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,256,0.0034835553831524322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,1024,0.019518221418062847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,6144,0.011878222227096558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,10240,0.019768888751665752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,16384,0.019907555646366544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,12288,0.2500480016072591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,10240,0.5502657890319824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,51200,0.10519555542204116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,2048,0.027427555786238775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,256,0.01051644484202067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,16384,0.12511821587880453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,64,0.023152889476882085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,10240,0.013547555440002017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,1024,0.008466666771305932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,3072,0.0857235524389479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,1024,0.023746665981080797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,16384,0.010175111393133799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,12288,0.19975554943084717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,128,0.006301333506902059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,3072,0.00795111142926746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,2560,0.2703999943203396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,2048,0.006484444356626934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,1024,0.010476444330480365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,3584,0.009451555709044138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,256,0.008043555749787224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,2048,0.012226666841242047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,16384,0.07561510801315308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,10240,0.6711226569281684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,1536,0.0074142225914531285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,256,0.04435555471314324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,16384,2.315784878200955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,65536,5.721763610839844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,256,0.02234933276971181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,2048,0.006157333238257303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,8192,0.01760444376203749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,64,0.0031066667288541794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,5120,0.06154400110244751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,3072,0.027753776974148218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,2560,0.5041226810879177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,1536,0.045216888189315796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,7168,0.026779555612140234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,32,0.003024000053604444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,1024,0.023843554986847773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,65536,0.11873688962724473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,256,0.0038773334688610504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,512,0.26042577955457896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,3072,0.04804977774620056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,65536,6.758394877115886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,4096,0.010301333334710863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,3072,0.015240000353919135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,12288,0.06274666388829549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,12288,0.06484266784456041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,3584,0.06995555427339342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,64,0.04836977852715386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,128,0.005508444375462002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,1536,0.03000266684426202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,65536,0.1286488903893365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,8192,0.007197333706749811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,8192,0.054006222221586436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,8192,0.6969191233317057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2048,768,0.06587200032340156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,768,0.017700443665186565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,2048,0.006065777606434292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,1536,0.5835093392266167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,6144,1.2754426532321506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,3584,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3584,2560,0.10888622204462688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,1536,1.3339164521959093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,2560,0.1923546658621894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,8192,0.010718222293588849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,1024,0.005859555469618903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,16384,0.13060088952382407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,128,0.005246222019195557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,6144,0.014192889134089151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,65536,0.5649253527323405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,16384,0.0358497765329149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,3072,0.01780533293883006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,3072,0.012356444365448423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,65536,0.11241243945227729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,3072,0.10456711053848267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,256,0.157167116800944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,5120,0.1674542162153456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,512,0.0038000001675552795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,5120,0.030905776553683813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,7168,0.05496266815397474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,51200,0.009659555223253038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,12288,0.043896890348858304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,51200,1.5295786327785914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,5120,0.03899288839764065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,512,0.018280888597170513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,64,0.002554666664865282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,3072,0.010081777969996134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,512,0.01536266671286689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,2560,0.0162453336848153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,256,0.005607111172543631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,7168,0.1260062191221449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,12288,0.015015999476114908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,3584,0.07202221949895223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,32,0.1859297752380371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,5120,0.21260978115929496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,12288,0.028024001253975764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,8192,0.02107644412252638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,1024,0.016725333200560678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,768,0.005680000202523337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,512,0.05576355589760674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,1024,0.20266666677263048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,7168,0.031215111414591473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,2048,0.011056000159846412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,51200,0.2972319920857747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,3584,0.03629688753022088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,128,0.00286666676402092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,16384,0.28610044055514866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,512,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,3072,0.02055111030737559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,3584,0.03471555643611484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,256,0.005531555662552516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,3072,0.008221333225568136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,32,0.005111999809741974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,65536,80.45348442925346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,1536,0.00886577781703737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,512,0.015557333827018738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,8192,0.12371377150217693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,64,0.003145777723855443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,65536,1.6213564342922633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,128,0.003080000066094928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,51200,0.48036977979871964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,10240,0.6902657614813911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,256,0.26182754834493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,4096,0.012879999975363413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,51200,0.32328799035814076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,256,0.036093334356943764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,3072,0.061900443500942655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,1536,0.0052488889131281115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,2048,0.03349599904484219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,6144,0.02944177720281813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1024,2560,0.01791111131509145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,8192,0.024288000331984624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,65536,3.300956514146593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,8192,0.045089777972963124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,768,0.019304000669055514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,4096,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,2048,0.04942844311396281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,6144,0.006635555376609166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,16384,3.5192612542046438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,1536,0.017651556266678702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,64,0.03430666526158651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,64,0.0033937777496046494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,16384,0.1042791141404046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,768,0.0258568889564938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,3072,0.3249128924475776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,2560,0.06245333618587918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,1024,0.027441778116756018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,10240,0.06052800019582113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,128,0.012456888953844706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,10240,0.050055109792285495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,2048,0.014857777290874057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,2560,0.011581333147154914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,8192,0.0209840006298489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,1536,0.004553777890072929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,512,0.006830222076839871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,32,0.00518222235971027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,2048,0.014520888527234396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,768,0.017090666625234816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,1024,0.10605866379208034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,16384,0.1367137829462687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,256,0.015591111448076038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,3584,0.053391112221611865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,5120,0.018444443742434185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,3072,0.047689778937233814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,1024,0.0063884444534778595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,2560,0.009594666461149851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,16384,0.023789332972632513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,2048,0.04138933287726508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,256,0.007487110793590546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,10240,0.017050666941536795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,10240,1.0025670793321397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,65536,0.1209297842449612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,4096,0.12641333209143743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,128,0.013857777747843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,2048,0.11077511310577393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,3072,0.06890933381186591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,8192,0.037189334630966187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,3072,0.01331022216214074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,2560,0.030611554781595867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,64,5120,0.011339555184046427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,2048,1.0584142473008897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,5120,0.22713510195414224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,65536,0.4119822184244792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,8192,1.630848036872016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,128,0.003568888952334722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,512,0.3017395602332221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,128,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,32,0.0026506665680143568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,768,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,2048,0.036630223194758095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,64,0.0027511111564106415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,768,0.18892178270551893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,2048,0.008184000021881526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,64,0.0069919998447100324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,128,0.005138666679461797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,2560,0.007167999943097432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,512,0.0033208889265855155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,1536,0.02459555533197191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,512,0.03212799959712558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,3584,0.008815111385451423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,6144,0.027288888891537983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,64,0.0030337776988744736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,4096,0.1323795583513048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,51200,0.0648346675766839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,1536,0.5794755617777506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,2560,0.00889155517021815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,3584,0.03285955720477634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,8192,0.010253333383136326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,51200,0.18633333841959634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,512,0.020592000749376085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,128,0.029846221208572388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,1024,0.004373333520359463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,768,0.005695999910434087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,1024,0.0051857779423395796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,16384,0.06247644291983711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,32,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,1536,0.005639111002286275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,768,0.005962666538026597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,64,0.005594666633341048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,51200,31.160101996527775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,16384,0.04858044452137417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,6144,0.009938666390048133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,256,0.005214222189452913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,32,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,3584,0.010439111126793755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,12288,0.02690222197108799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,65536,0.21104266908433703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,3584,0.014069333672523499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,7168,0.022475555539131165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,12288,1.0104604297214084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,768,0.007605333295133378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,1024,0.04717688759167989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,768,0.005506666584147348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,1024,0.0068195557428730865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,3584,0.03438755538728502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,32,0.1575360033247206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,65536,0.2979493406083849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,3584,0.015788444214397006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,64,0.011591110792424945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,128,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,32,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,51200,0.3241048918830024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,64,0.008232889076073965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,1024,0.013785777820481194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,512,0.02257866660753886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,12288,0.4928080240885417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,2560,0.031727999448776245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,3584,0.02941511074701945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,1024,0.005791999813583162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,12288,0.037194665935304426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,65536,1.3923457463582356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,768,0.006218666831652324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,10240,0.27435554398430717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,10240,0.017831110292010836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,1024,0.030827555391523574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,512,0.26961421966552734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,16384,0.013579555683665805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,512,0.003564444267087512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,3584,0.02854222059249878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,768,0.003963555726740095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,3584,0.009120889008045197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,6144,0.02762577765517765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,4096,0.008075555165608725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,256,0.015824889143308003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,128,0.004912888838185204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,7168,0.13968177636464438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,16384,0.21703733338250053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,1024,0.016207999653286405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,2560,0.20264977878994414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,2048,0.021228444245126512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,1536,0.010838222172525195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,128,0.0046311111913787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,3072,0.02182666626241472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,10240,0.1478417714436849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,2048,0.11096533139546712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,1536,0.006233777850866318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,128,0.004328000048796336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,1024,0.006649777707126405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,64,0.014261333478821648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,7168,0.2618809011247423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,7168,0.034352001216676496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,1536,0.49878401226467556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,12288,0.009313777916961247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,6144,0.619465774959988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,4096,0.014596444037225513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,16384,0.7809742291768392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,16384,0.41084533267551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,512,0.013634666800498962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,10240,0.019514666663275827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,5120,0.18217689461178252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,3072,0.032440000110202365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,1536,0.010075555907355415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,12288,0.011048888994587792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,2048,0.005323555734422471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,6144,0.012303111453851065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,4096,0.020015999674797058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,4096,0.019076445036464267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,1536,0.07624444696638319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,128,0.005728888842794631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,8192,0.2425395647684733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,8192,0.03825066818131341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,7168,0.09609066777759129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,5120,0.4235528839959039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,64,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,32,0.005513777749405966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,1536,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,3072,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,7168,0.31916711065504283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,3584,0.085043556160397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,4096,0.15193955103556314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,768,0.004843555390834808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,51200,0.07130844725502862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,128,0.003543111185232798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,1536,0.10216799709531997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,3072,0.007724444071451823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,768,0.0041164445380369825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,6144,0.06460089153713651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,16384,128,0.0096142225795322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,51200,15.044152153862846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,65536,0.14879822731018066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,8192,0.32176534334818524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,3584,0.009639110830095079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,65536,0.7571155760023328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,1536,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,7168,0.021664000219768945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,3072,0.0454657773176829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,16384,0.10371733374065822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,2560,0.009156444834338294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,1536,0.5441004435221354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,10240,0.22910221417744955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,3584,0.06885066959593031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,16384,0.05184266633457608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,7168,0.015184000134468079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,64,0.026826666461096868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,4096,0.014121777481502958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,4096,0.3512595494588216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,128,0.00794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,12288,0.05623200204637316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,12288,7.304899427625869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,3584,0.012853333519564735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,51200,2.9392435285780163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,1024,0.04050666756100125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,128,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,256,0.009120889008045197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,128,0.004514666481150521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,16384,0.07723644706938002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,256,0.0037395556767781577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,7168,0.013329777452680798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,5120,0.07012533479266696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,65536,3.0081475575764975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,10240,0.15795021586947972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,5120,0.019951999187469482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,1024,0.004634666774008009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,256,0.014117333624098035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,6144,0.06222044759326511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,512,0.003149333306484752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,32,0.009299555586444007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,32,0.0025519999778932994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,65536,0.06511555777655707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,512,0.010377777947319878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,512,0.01882577770286136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,2048,0.015087111128701104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,128,0.006818666640255187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,3584,0.011148444480366178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,256,0.004242666479614046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,3072,0.006351999938488007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,768,0.006120000034570694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,3072,0.03502488798565335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,1536,0.018600889378123812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,3584,0.019126221537590027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,12288,1.201218711005317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,6144,0.11568266815609402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,3584,13.235509236653646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,7168,0.026154667139053345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,4096,0.017009778155220878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,6144,0.09502666526370579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,3072,0.06658399767345853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,1536,0.0734764469994439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,2560,0.012619555824332766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,8192,0.05098488926887512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,4096,0.013643555343151093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,1024,0.03497511148452759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,16384,3.9642177157931857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,7168,0.06519555383258395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,4096,0.4423191017574734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,65536,6.417784796820746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,256,0.00923111124171151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,1024,0.0129120002190272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,65536,0.513214217291938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,65536,0.06090488698747423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,65536,0.13932355244954428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,768,0.006474666711356904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,128,0.010965333216720156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,512,0.009111111362775167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,2560,0.012274666792816587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,2048,0.04876710971196493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,256,0.007878222399287755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,6144,0.08131288819842868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,4096,0.02082577844460805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,1536,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,512,0.014682667122946845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,768,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,8192,0.035014222065607704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,64,0.007309333317809635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,6144,0.009717333647939894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,128,0.002842666581273079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,3584,0.03157777918709649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,2560,0.12024622493320042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,2048,0.006618666566080517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,51200,0.6616124576992458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,6144,0.015434667468070984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,5120,0.025018667181332905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,64,0.004046222401989831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,1536,0.0070497774415545994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,6144,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,768,0.005222222043408289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,64,0.014460444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,16384,0.9527031580607096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,128,0.024463999602529738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,12288,0.2186408837636312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,2560,0.013608000344700284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,16384,0.0236835562520557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,51200,0.062365333239237465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,65536,0.2788684368133545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,1024,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,51200,0.21918843852149117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,2048,0.01347733371787601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,3584,0.017064889272054035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,6144,0.018014222383499146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,3584,0.04858933223618401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,8192,0.017649778061442904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,4096,0.0609146687719557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,7168,0.0910595522986518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,7168,0.06822577450010511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,2560,0.00793688909875022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,128,0.003165333428316646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,65536,0.9921991560194227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,12288,0.20523556073506674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,2560,0.557695123884413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,7168,0.02165777815712823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,256,0.04275111026234097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,512,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,4096,0.014850666125615438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,3584,0.32548711034986705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,6144,0.005972444597217772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,768,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,4096,0.010207110808955299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,7168,0.013779555757840475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,5120,0.008236444658703273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,128,0.009391999906963771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,64,0.03581600056754218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,64,0.003567999849716822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,12288,0.023066666391160753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,128,0.09293688668145074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,5120,0.04343200061056349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,256,0.007170666423108842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,65536,2.5502329932318792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,4096,0.012572444147533841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,3072,0.05936800109015571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,65536,0.09051911036173503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,3072,0.00870488915178511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,12288,0.08505333132214016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,768,0.004010666575696734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,51200,0.1455448865890503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,16384,0.024523556232452393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,2560,0.007688889073001013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,768,0.006251555350091722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,1024,0.007852444218264686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,2048,0.05050488975312975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,6144,0.016511999898486666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,256,0.006276444428496891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,3584,0.04877333177460564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,256,0.008188444707128737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,5120,0.03569599986076355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,32,0.0052791109515561005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,8192,0.02911111050181919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,128,0.009512888888518015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,65536,0.01314222233162986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,256,0.13350755638546413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,65536,0.17821510632832846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,4096,0.18681866592831084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,2560,0.03786844346258376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,4096,0.010455110834704505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,7168,0.015967999895413715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,1024,0.056162668599022754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,6144,0.051925334665510386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,8192,0.006206222292449739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,2560,0.013172444370057849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,10240,1024,0.0092284447617001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,128,0.0047591109242704184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,6144,0.01570399933391147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,512,0.005571555760171678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,12288,0.0973128875096639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,768,0.02352977792421977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,8192,0.01367111090156767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,768,0.011056889262464313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,65536,1.5360693401760523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,6144,0.07327644692526923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,4096,0.024215110474162634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,128,0.04418666826354133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,51200,0.05398577782842848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,3072,1.0717439651489258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,4096,0.03817777832349142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,2048,0.00573600000805325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,8192,0.08204444249471028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,4096,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,10240,0.4873510996500651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,512,0.003825777934657203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,3584,0.015587555037604438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,10240,0.06752977768580119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,2560,0.022856000396940444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,65536,0.18835911485883924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,256,0.011063999599880643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,3072,1.3909048504299586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,51200,0.07538311349021064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,65536,0.14968000517951116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,10240,0.05863910913467407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,512,0.005723555468850666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,3584,0.05079288946257698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,3584,0.011528888510333167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,10240,0.03315466642379761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,8192,0.014709333578745524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,2560,0.011585777832402123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,10240,0.023073777556419373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,5120,0.010854221880435944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,64,0.003628444340493944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,65536,0.7882506582472059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,1024,0.009259555902745988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,16384,0.15000532733069524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,65536,0.2615404393937853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,12288,0.06677866644329496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,64,0.004067555483844546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,3584,0.01789688898457421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,51200,0.03459111187193129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,1024,0.015226667126019796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,7168,0.010785777535703448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,7168,0.03056266572740343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,1024,0.0058311112225055695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,1536,0.10439199871487087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,3072,0.010051555931568146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,128,0.0032684444967243406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,512,0.02517066730393304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,32,0.006529777828190062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,768,0.0035982223020659555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,6144,64,0.034909334447648786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,16384,0.6889866722954644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,12288,0.510091569688585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,4096,0.00654755574133661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,1536,0.012764444781674279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,65536,0.1337991158167521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,6144,1.9883573320176866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,1024,0.0075235557225015424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,65536,0.1492746671040853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,16384,0.03163733416133457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,5120,0.09895022047890557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,1536,0.01056177748574151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,6144,10.25982920328776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,512,0.039684444665908813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,256,0.032998220788107976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,768,0.020170667105250888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,12288,0.04718755682309469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,7168,0.04136177897453308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,1536,0.014024888475735983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,4096,0.15822577476501465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,256,0.016135111451148987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,3072,0.01165244479974111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,1024,0.034267554680506386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,12288,0.023609777291615803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,2560,0.025123554799291823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,1024,0.0058871110280354815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,65536,0.8548906644185384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,65536,0.3419271045260959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,2560,0.006113777971929974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,256,0.013720889058378009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,3072,0.051832887861463756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,10240,0.06709066364500257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,10240,1.134804407755534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,65536,0.8473759757147895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,512,0.004791999856630961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,10240,0.2740817864735921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,4096,0.011722666521867117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,65536,0.6055955357021755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,3072,0.011368888947698804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,6144,0.014316444595654806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,1536,0.008452444440788692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,16384,0.04704799917009142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,10240,0.052721778551737465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,256,0.003173333282272021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,16384,0.013968888256284924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,5120,0.03995910949177212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,8192,0.8810808923509387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,8192,0.1525084442562527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,10240,0.1525768836339315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,7168,0.32126932673984104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,128,0.007334222396214803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,12288,0.3158835570017497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,256,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,5120,0.04750222298834059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,16384,256,0.039862222141689725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,2560,0.00812177773978975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,64,0.005245333330498801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,7168,0.01089599976936976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,768,0.16581332683563232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,768,0.00600355573826366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,5120,0.14324177636040583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,7168,0.7404310968187121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,768,0.006467555546098285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,768,0.01000533335738712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,12288,0.1470302210913764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,1024,0.007425777614116669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,256,0.0042275554604000514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,512,0.020784000555674236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,51200,0.6671368810865613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,51200,0.8780106438530816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,10240,0.06357777780956693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,1024,0.004724444614516364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,51200,5.005298614501953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,12288,0.048638221290376454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,2048,0.06237510840098063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,2048,0.16654488775465223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,2560,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,768,0.007384888827800751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,2560,0.6598773532443577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,12288,0.16179111268785265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,1024,0.009655999640623728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,51200,0.02595111065440708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,4096,0.011984888878133563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,4096,0.008043555749787224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,16384,0.10383288727866279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,128,128,0.00674222202764617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,10240,0.014023999373118082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,8192,0.014880888991885714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,64,0.0029519999192820657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,768,0.005417777846256892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,512,0.005392888767851724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,65536,0.019104000594880845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,128,0.0041128889554076726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,512,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,512,0.00350222239891688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,8192,0.027386666999922857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,5120,0.050513777467939586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,16384,5.4567820231119795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,51200,0.745680914984809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,1536,0.010115555591053432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,4096,0.21837777561611602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,3072,0.06432177623112996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,16384,0.0550871127181583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,16384,0.03938044442070855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,2048,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,768,0.025721778472264607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,3584,0.02272355556488037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,1536,0.007012444237867991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,256,0.011790222591824003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,512,10240,0.08246755599975586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,5120,2048,0.00983199973901113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,8192,0.044363554981019765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,51200,0.7474213176303439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,5120,0.008481777376598781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,64,0.025195555554495916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,65536,0.6403999858432347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,16384,0.03334310981962416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,128,0.020356444848908316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,1024,0.025969778498013813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,512,0.29532088173760307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,1024,0.01279022213485506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,5120,0.011752000285519494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,512,0.02857066525353326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,4096,0.013752000199423896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,7168,4.816371493869357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,12288,0.09892266988754272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,512,0.019730667273203533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,8192,0.4171831077999539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,2048,0.06803466876347859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,8192,0.014180445008807711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,64,0.03198488884501987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,2560,0.03552444444762336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,51200,0.2917439937591553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,16384,0.007672888537247975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,8192,0.03232444326082865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,16384,8.151236640082464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,10240,0.12179022365146214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,128,0.027462222509913977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,256,0.006717333363162146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,16384,0.0664079984029134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,256,0.0064444442590077715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,16384,0.058261334896087646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,8192,0.14047466384039983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,51200,0.11387733618418376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,768,0.017265778448846605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,3072,0.16513333055708143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,3072,0.00654666663871871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,51200,2.363135019938151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,128,0.033152000771628484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,1536,0.025451555848121643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,1024,0.09202666415108575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,7168,51200,0.12747022840711805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,4096,0.016530666086408827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,3584,0.029254221253924902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,16384,0.0993644462691413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,4096,0.011232888533009423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,5120,0.013920888304710388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,5120,0.01891288823551602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,2560,0.11883644262949626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,512,0.03222222129503886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,4096,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,512,0.006194666855865055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,1536,0.005700444595681296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,512,0.009409777820110321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,64,0.007816889219813878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,4096,0.012200888660218982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,3584,0.01566133399804433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,7168,0.007460444337791867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,10240,0.013739555246300168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,8192,0.2114586697684394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,1536,0.07521422041787042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,3072,0.03021777669588725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,4096,0.007785778078767989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,2560,0.04067822297414144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,51200,1.08695560031467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,5120,0.030836443106333416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,8192,0.09934044546551174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,1024,0.0046844445168972015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,16384,0.09115911192364162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,3584,0.02683999968899621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,8192,0.5164933204650879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,8192,0.032681776417626276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,4096,0.011218667030334473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,32,0.018938667244381376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,65536,10240,0.5346408949957954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,8192,0.015560888581805758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,5120,0.2123653358883328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,32,0.0028533333291610083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,4096,0.00927111092540953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,1024,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,16384,0.024004444479942322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,768,0.007928888830873702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,64,0.1006559994485643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,1536,0.006344888773229387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,128,0.02905333373281691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,4096,0.008739555875460306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,3072,0.01901333365175459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,6144,0.05411022239261203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,1536,0.02325422234005398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,6144,0.03201777736345927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,5120,0.032947556840048894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,8192,0.14530932903289795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,10240,0.009370666411187913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,1024,0.005696889013051987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,3072,0.008895999855465358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,7168,0.02411199940575494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,10240,0.013253333667914072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,7168,0.01128533316983117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,256,0.0063928887248039246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,4096,0.015027556154463025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,65536,0.5342897839016384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,7168,0.07316444317499797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,10240,36.24120754665798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,3072,0.011734222372372946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,768,128,0.002874666617976295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,7168,0.04617422156863742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,16384,0.06715289089414808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,6144,0.09772888819376628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,32,0.010009778042634329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,2560,0.20556978384653726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,8192,0.07082044416003756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1536,65536,0.07612355550130208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,6144,0.02293511066171858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,3072,0.014630221658282809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,3584,0.00945688866906696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,32,0.003126222226354811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,64,0.016494222813182406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,5120,0.009645333720578088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,16384,1.493393792046441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,1024,0.034712887472576566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,256,2048,0.030105776256985132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,64,0.00535733335547977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,32,0.0027911110470692315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,2048,0.10712977912690903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,768,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,16384,0.023955555425749883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,12288,0.3477075629764133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,1024,0.01254755589697096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,65536,0.4380204412672255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,2560,0.019117333822780185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,2048,0.027848889430363972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,8192,0.025976000560654536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,5120,0.008480889101823172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,7168,0.04967911044756571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,3584,0.023870221442646448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,768,0.0038373333712418876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,2048,0.006958222223652734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,2048,0.007233777807818518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,512,0.0035146665241983202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,3072,0.015878222054905362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,1536,0.009870222045315636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,256,0.004296888907750447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,8192,0.04689155684577095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,128,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3584,1024,0.025549333956506517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,10240,0.006694222076071634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,512,0.006159111029571957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,65536,0.20383911662631563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,7168,0.10957422521379258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,2048,0.0053226666318045715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,768,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,3072,0.014135999812020196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,128,0.008555555509196388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,256,0.004825777891609403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,2048,0.0355022218492296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,32,0.00508355556262864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,3584,0.05463555455207825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,4096,0.0110880004035102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,256,0.007425777614116669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,16384,0.04718666606479221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,16384,0.029192888074451025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,8192,0.02667288813326094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,7168,0.05548000004556444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,512,0.008902221918106079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,65536,2.6710809071858725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,7168,1536,0.07760710848702325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,51200,0.0903315544128418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,5120,0.2611893282996284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,10240,0.11116622553931342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,6144,0.041402667760849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,3584,0.01129066695769628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,1536,0.017838221457269456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,1536,0.008106666306654612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,8192,0.02221155497762892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,1024,0.04972088999218411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,1024,0.017663111289342243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,7168,0.010859555668301053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,512,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,3072,2.2764053344726562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,512,0.005056888692908817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,3072,0.2274035612742106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,12288,0.022662222385406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,65536,0.07960444688796997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,2048,0.06870666477415296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,4096,0.3159128824869792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,256,0.0037128888070583344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,5120,0.093941330909729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,64,0.00425511101881663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,12288,0.019511111908488803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,1024,0.03277422322167291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,4096,0.28589243359035915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,512,0.012655110822783576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,64,0.0046924443708525766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,2048,0.013139555851618448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,512,0.008551110823949179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,768,5120,0.009367110828558603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,128,0.007992888490358988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,4096,0.006731555693679386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,8192,0.006471111128727595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,16384,0.13923554950290257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,1024,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,5120,1.2559102376302083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,64,0.0025742221623659134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,64,0.004975999808973736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,7168,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,32,0.006801777829726537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,256,0.03654133280118307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,512,0.010558221903112201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,32,0.0027626665929953256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,256,0.004671999977694618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,64,0.008423999779754216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,32,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,6144,0.10030488835440741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,768,0.009872888525327047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,1536,0.008037333687146505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,3584,0.00960355583164427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1536,4096,0.052721778551737465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,12288,0.12192889054616292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,768,0.004217777815130022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,3584,0.23530133565266928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,3072,0.018780445059140522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,4096,0.014241778188281588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,7168,0.012980444563759698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,256,0.024684444069862366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,3072,0.02774222195148468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,65536,0.9511519538031684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,12288,0.038513776328828596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,5120,0.019499555230140686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,51200,4.340870327419705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,12288,0.062056885825263135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,6144,0.1699111064275106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,64,0.01663644446267022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,4096,0.08067466815312703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,3072,0.24943378236558703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,3072,0.014055111342006259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,65536,6.643512725830078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,256,0.006804444309737947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,16384,0.8364222314622668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,4096,0.023215111758973863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,6144,0.010949333508809408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,3072,0.12767910957336426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,12288,0.036575112077924944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,32,0.01127022256453832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,768,0.060263997978634305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,8192,0.1509351068072849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,3584,0.010406222608354356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,1024,0.03593333231078254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,32,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,51200,0.14919022719065347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,256,0.012526222401195101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,1024,0.005078222188684675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,3584,0.009876444107956356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,51200,0.09649866819381714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,65536,1.2120960023668077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,6144,0.24724088774787056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,1024,0.45243464575873477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,128,0.0063928887248039246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,3584,0.007824888659848107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,4096,0.03048444456524319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,5120,0.042187554968727954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,6144,0.13738755385080972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,512,0.0058328890138202244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,51200,0.2726355658637153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,256,0.014558222558763293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,8192,0.04839288857248095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,1024,0.005706666658322017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,128,0.005374222166008419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,512,0.01774133410718706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,1536,0.005530666559934616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,6144,0.09297777546776666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,128,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,64,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,3072,0.03862577676773071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,51200,4.907470279269748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,1024,0.0111191107167138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,3584,0.015641777051819693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,3584,0.05162044366200765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,7168,0.017620444297790527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,1536,0.010473777850468954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,5120,0.03747999999258254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,10240,0.4512302080790202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,2048,0.007681777907742395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,12288,0.040728890233569674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,8192,0.006813333266311222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,4096,0.00832444429397583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,32,0.003059555465976397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,1536,0.004902222090297275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,3584,0.2661244339413113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,10240,0.03633866707483927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,2560,0.024744888146718342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,10240,0.021457778082953557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,256,0.01074577785200543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3072,10240,0.047720889250437416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,10240,0.36051021681891543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,3072,0.008260444634490544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,7168,0.02327288852797614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,64,0.005843555761708154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,3072,0.011200889117187925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,6144,0.13869510756598577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,128,0.007917332980367873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,51200,0.7344195577833387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,12288,0.013928888572586907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,51200,0.19864711496565077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,2048,0.04014755619896783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,128,0.012474666866991254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,32,0.004491555607981152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,256,0.019021333919631112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,3072,0.011188444164064197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,1536,0.014352000421947904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,768,0.02910400099224514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,12288,5.775941212972005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,6144,0.48334042231241864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,768,0.004479111068778568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,2048,0.014225777652528552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,64,0.0038684445122877755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,65536,3.072051578097873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,51200,0.11780444780985515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,4096,0.16952266958024767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,64,0.2334053251478407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,1024,0.02254044512907664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,768,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,65536,0.18274577458699545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,128,0.005098666581842634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,3584,0.07094133562511869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,768,0.0192248887485928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,256,0.01017599966790941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,65536,1.2238746219211154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,4096,0.07526310947206286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,6144,0.007999999655617608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,51200,0.16966756184895834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,768,0.0058337777025169795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,8192,0.02456088860829671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,128,0.0028515555378463534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,16384,0.06193421946631538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,2048,0.08480266729990642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,64,0.0026133333643277488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,5120,0.043613334496816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,768,0.008851555486520132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,2560,0.02681066592534383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,51200,0.05876355701022678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,1536,0.012287110918098025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,2560,0.009894222021102905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,16384,0.0211191111140781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,128,0.003072888900836309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,32,0.0026746667507621977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,7168,0.019699555304315355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,2560,0.023223999473783705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,51200,1.4856897989908855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,256,0.011869333684444427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,7168,0.053101334306928843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,12288,0.8842106925116645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,32,0.006545777950021956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,128,0.03667466508017646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,2560,0.09879644711812337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,65536,0.2606995635562473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,12288,0.18078666263156465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,3072,0.0513644450240665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,256,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,12288,0.011930666863918304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,4096,0.27352621820237905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,8192,0.07236088646782769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,51200,0.10408266385396321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,6144,0.015995555453830294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,12288,0.33898578749762637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,512,0.03133599956830343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,64,0.010980444649855295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,32,0.010999110837777456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,8192,0.6188195546468099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,1536,0.04054488738377889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,3072,0.008163555628723567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,32,0.0069644442862934535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,12288,0.10293689039018418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,10240,0.03363110952907138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,3584,0.011800888511869641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,2560,0.04843911197450426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,51200,1.9415946536593969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,16384,0.03137333194414774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,256,0.0034773333205117118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,16384,2048,0.23799199528164336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,2560,0.008637333081828224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,128,0.004963555683692296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,3072,0.12813156180911595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,5120,0.1713200012842814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,512,0.004755555755562252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,1536,0.010002666877375709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,2048,0.055997331937154136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,16384,0.020185778538386028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,6144,0.02388800018363529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,3584,0.020572443803151447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,3584,0.017400888933075797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,512,0.003914666672547658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,51200,0.012730666332774691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,5120,0.019182221757041085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,2560,0.00602311102880372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,6144,0.1730417807896932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,256,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,768,0.00360355567600992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,16384,1.6295341915554469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,1536,0.48144891526963973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,7168,0.06730222039752536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,2048,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,3072,0.016512889001104567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,2560,0.009639999932712978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,8192,0.035747554567125105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,32,0.0029297777348094513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,64,0.013090666797426013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,65536,0.918659528096517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,16384,0.028251555230882432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,768,0.005923555543025334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,2560,0.008135110967689091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,12288,0.17344533072577584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,256,0.01367555558681488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,768,0.011752000285519494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,3072,0.2016293340259128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,256,0.004648000001907349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,512,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,3584,0.23278933101230195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,768,0.043056888712777026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,10240,0.01386755539311303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,128,0.00600888869828648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,12288,0.6237182087368435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,12288,0.08371910783979629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,64,0.003010666618744532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,2560,0.01601777805222405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,32,1024,0.00599644457300504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,8192,0.09040000041325887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,16384,0.04874044325616625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,2048,1.1815368864271376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,1536,0.011525332927703857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,256,0.006503110958470239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,1536,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,64,0.008808000220192803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,7168,0.7541715304056803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,4096,0.12473244137234157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,3584,0.1215475532743666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,128,0.0052977779673205475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,256,0.004212444441186057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,1536,0.1475813388824463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,2560,0.008936000367005667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,12288,0.01517955627706316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,5120,0.021856000026067097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,128,0.024152888192070857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,32,0.002593777866827117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,3584,0.10516444179746841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,12288,0.021631111701329548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,16384,0.012321777641773224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,51200,0.013070222404268054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,12288,0.019998222589492798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,2560,0.019055111540688407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,512,0.0038364442686239877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,5120,0.012512889173295764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,8192,0.023355555203225877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,65536,0.36554577615525985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,768,0.13263644112481013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,7168,0.2957342200809055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,8192,0.12479910585615371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,2048,0.06465333037906222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,256,0.0031679999083280563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,3584,0.015580443872345818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,7168,0.014396443963050842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,4096,0.22065511014726427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,6144,0.03916088740030924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,3072,0.15031911267174616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,16384,0.6085999806722006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,1536,0.009272000028027428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,12288,0.5235439936319987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,1536,0.009191999832789103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,16384,0.009568000005351173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,1024,0.0873839987648858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,2048,2.917734146118164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,8192,0.06101510922114054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,7168,0.009122666385438707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,65536,0.20860889222886828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,512,0.0069751110341813825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,6144,0.014871110518773397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,10240,0.026480888326962788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,12288,0.06259022156397502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,3584,0.03843911157713996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,1024,0.004559111264016894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,2560,0.006858666737874349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,16384,0.40679465399848086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,256,0.005416000054942236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,256,0.01612444387541877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,3584,0.020593777298927307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,6144,0.025573333104451496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,256,0.023843554986847773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,256,0.008738666772842407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,65536,1.8888123830159504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,64,0.0028640000770489373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,128,0.0029271110478374693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,65536,0.6504062016805013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,512,0.023022222850057814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,10240,0.7218640115525988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,1024,0.02843377656406826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,2560,0.05585155884424845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,4096,0.05728622277577718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,2560,0.36938932206895614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,2560,0.3161279890272352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,512,0.0034942221310403612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,16384,0.07600266403622098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,65536,0.1466968854268392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,3072,0.008550222549173567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,768,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,5120,0.018585777944988675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,512,0.003084444544381566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,256,0.0047128887640105355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,65536,0.08520355489518906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,256,0.0033359999457995095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,8192,0.01278222186697854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,5120,0.14223644468519422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,4096,0.00795555528667238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,1536,0.0123511114054256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,2560,0.030235555436876085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,8192,0.021479111578729417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,4096,0.028357333607143823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,2048,0.010023111270533668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,32,0.005391110976537068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,3072,0.00962488849957784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,1024,0.25177955627441406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,8192,0.019907555646366544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,8192,16384,0.046394666035970054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,128,0.005144000053405762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,6144,0.01975822283162011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,128,0.005648889061477449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,3072,0.014764444695578681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,6144,0.012141333685980903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,10240,0.021328889661365088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,2048,0.07724089092678495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,4096,0.1787173350652059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,3072,0.01994311147265964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,2560,0.04315111041069031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,4096,0.010842666857772403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,10240,0.020051555501090158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,10240,0.21862222088707817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,3584,0.0063484443558586975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,2048,0.0409786668088701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,1536,0.027895109521018133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,3072,0.3026062117682563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,12288,0.05403466688262092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,128,0.0027031112048361036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,32,0.0069217777086628815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,32,0.007552000383536021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,65536,0.15153333875868055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,4096,0.06017866399553087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,256,0.004843555390834808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,12288,0.31560177273220485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,256,0.010619555910428366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,512,0.03650133477316962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,6144,0.10899200042088826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,3584,0.02741866641574436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,2048,0.008005333443482717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,12288,0.04018222292264303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,51200,0.6621484226650661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,2048,0.7903075748019748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,2560,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,64,0.004605333424276776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,3072,0.015404444601800708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,12288,0.05710666709476047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,2560,0.14972533120049372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,8192,0.017248888810475666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,8192,0.10642488797505696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,5120,0.020608888732062448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,8192,0.3567039966583252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,768,0.0270506673389011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,12288,0.0417520006497701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,128,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,16384,2.550850762261285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,256,0.005127999931573868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,512,0.0321760012043847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,10240,0.01681155628628201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,12288,0.010015111002657149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,10240,0.027270222703615826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,6144,0.26573067241244847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,2048,0.011173333558771344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,5120,0.07348089085684882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,51200,0.15448799398210314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,7168,0.13562577300601536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,2048,0.9706524742974175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,256,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,3584,3.145668453640408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,10240,0.008403555386596257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,512,0.016872889465755887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,10240,0.008443555898136562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,6144,0.013197333448463015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,3072,0.01092888911565145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,10240,0.07287466526031494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,1536,0.06907644536760119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,128,0.008058666355080074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,1024,0.007039111521508958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,65536,0.25865689913431805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,3584,0.012251555919647217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,1536,0.012797333300113678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,12288,0.03190933333502875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,2048,0.017250667015711468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,128,0.005425777700212266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,51200,0.2930026584201389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,65536,0.2220568921830919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,65536,0.3092711236741808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,256,0.0034737777378824023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,51200,1.6524115668402777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,10240,0.2948959933386909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,12288,0.09842577907774185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,6144,0.21467110845777723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,64,0.00300444434914324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,128,0.007450666692521837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,32,0.003104000041882197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,2560,0.038316445218192205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,3072,0.009365333451165093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,3072,0.3527982234954834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,16384,0.020814221766259935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,128,0.02643288837538825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,128,0.015320888823933072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,5120,0.09477688868840535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,768,0.09353955586751302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,128,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,10240,2048,0.012568000290128918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,4096,0.06207466787762112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,64,0.007048000064161088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,768,0.005712888720962737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,512,0.005295111073387994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,64,0.0026613333159022858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,64,0.002586666701568498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,6144,15.720542060004341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,512,0.24817334281073677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,4096,0.010580444501505958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,65536,0.09912888871298896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,2048,0.0321706665886773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,1024,0.11999200450049506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,65536,7.4159800211588545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,512,0.01551555593808492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,65536,0.861106660630968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,3584,0.01866933372285631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,512,0.2314088874393039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,6144,0.014675555957688225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,128,0.005723555468850666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,768,0.006671111202902264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,10240,0.014360000689824423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,16384,0.04977955420811971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,3584,0.05428266525268555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,128,0.029185778564876978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,5120,0.0790133343802558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,4096,0.009544889132181803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,4096,0.05256444546911451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,7168,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,2048,0.00664533343580034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,12288,0.21384621991051567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,3072,0.015276443627145557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,65536,0.45700624254014754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,65536,0.06953689124849108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,1024,0.47419198354085285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,7168,2.1577972835964627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,51200,0.149327105946011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,16384,0.09862399763531154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,32,0.010328888893127441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,128,0.009617778162161509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,12288,0.044340444935692676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,7168,0.014475555883513557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,3072,0.1062542200088501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,7168,0.28798309961954754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,2048,0.11871200137668186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,1536,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,512,0.008229333493444655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,1536,0.00814133303032981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,128,0.012320888539155325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,3584,0.1164968940946791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,7168,0.014549333188268872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,5120,0.008092444803979661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,7168,1.485280884636773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,5120,0.08158577813042535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,2048,0.007476444045702617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,16384,0.07084888882107206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,16384,0.011944889194435544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,4096,0.025438222620222304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,1536,0.010136889086829292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,5120,0.06471377611160278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,1024,0.040968888335757785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,3072,0.007431999676757389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,4096,0.17038399643368193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,5120,0.01882933411333296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,1024,0.011741333537631564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,3584,0.024013333850436743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,6144,0.02441155579355028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,5120,0.014380445082982382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,3072,0.047268443637424044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,512,0.04403377903832329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,768,0.010890666809346942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,768,0.09746755494011773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,128,65536,0.016726222303178575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,2048,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,8192,0.013831111292044321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,51200,0.3770471149020725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,768,0.048022223843468555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,4096,0.05033066537645128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,65536,0.10077866580751206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,1024,0.00480000012450748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,6144,0.01663466625743442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,3584,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,128,0.006710222197903528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,8192,30.8312750922309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,4096,0.014399111270904541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,7168,0.02830222249031067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,128,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,4096,0.10923644569185044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,768,0.011896889242861005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,12288,0.030125333203209773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,3584,0.13552267021603054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,2048,0.011049778097205691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,32,0.005773333211739858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3072,2560,0.09727022382948135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,32,0.01737866633468204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,6144,0.009860444400045607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,6144,0.012514666550689273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,12288,0.01888088881969452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,768,0.01834488908449809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,768,0.02571644385655721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,256,0.007546666595670912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,5120,0.2721253236134847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,1024,0.018453333112928603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,8192,0.03419999943839179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,3584,0.2416586611005995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,768,0.005555555638339784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,8192,0.06757866673999362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,64,0.008585777547624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,7168,0.13590489493476018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,512,0.12033600277370876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,64,0.007489778101444244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,7168,2.9125599331325955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,32,0.003370666669474708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,10240,0.03257333238919576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,128,0.09810400009155273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,2560,0.03603200117746989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,5120,0.015767110718621146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,6144,0.012711111042234631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,6144,0.02606577840116289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,6144,0.008572444319725037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,12288,0.3650391101837158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,16384,10.28078884548611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,10240,0.05305244525273641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,1024,0.016985777351591323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,1536,0.09266755315992568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,768,0.01219288922018475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,5120,0.04553066690762838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,3072,0.18465599748823378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,4096,0.06557244724697538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,512,0.018748443987634447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,1536,0.935058699713813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,64,0.003132444495956103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,128,0.01685599982738495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,7168,0.018266666266653273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,5120,0.012919999659061432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,2560,0.7587359746297201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,8192,0.016391999191708036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,128,0.005605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,1024,0.02894222074084812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,256,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,65536,0.10982133282555474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,12288,0.011863110793961419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,2560,0.014205333259370593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,256,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,64,0.009986666341622671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,16384,2.765896055433485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,128,0.021394666698243883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,6144,0.0927128858036465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,8192,0.1282151142756144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,6144,0.161354660987854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,1024,0.08218933476342095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,51200,0.5159982045491537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,16384,2.293257819281684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,128,0.029234667619069416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,128,0.0038035557501845886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,8192,2.5019609663221574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,1536,0.008366222182909647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,2560,0.010879110958841113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,2560,0.12584800190395778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,4096,0.05437511205673218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,51200,0.7725546624925402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,768,0.0036951113078329298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,10240,0.3940613269805908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,768,0.04832444588343302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,512,0.003759110967318217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,6144,0.017883555756674874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,1024,0.006873777757088344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,32,0.023523555861579046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,2048,0.055693335003323026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,12288,0.2049084504445394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,512,0.005411555369695027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,16384,1.006291601392958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,64,0.005952000204059813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,128,0.003539555602603488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,6144,0.07920889059702556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,4096,0.18443377812703451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,10240,65536,0.20162310865190294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,512,0.030377778742048476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,4096,0.022329777479171753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,2560,0.043567998541726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,6144,0.06798488563961454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,8192,0.01240533341964086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,6144,0.05253422260284424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,2560,0.018012444178263348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,51200,0.1233359972635905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,256,0.004229333251714706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,12288,0.0381368895371755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,32,0.0053484443989064955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,7168,0.06207466787762112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,8192,0.014098667436175875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,12288,0.0074106670088238185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,65536,41.86760457356771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,4096,0.5138319863213433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,32,0.022878222995334204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,1024,0.00886933339966668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,2560,0.016376000311639573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,10240,0.026937777797381084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,2560,10.349159240722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,1536,0.07660355832841662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,3584,0.058542225095960826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,64,0.03445422318246629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,10240,0.03971111112170749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,12288,2.163869433932834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,3584,0.6903182135687934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,2048,0.019896000623703003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,4096,0.013752889302041797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,768,0.09004177649815877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,512,0.04048177931043837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,32,0.0033688888781600525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,51200,0.18783022297753227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,4096,0.05532000131077237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,512,0.004092444562249714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,16384,0.013010666602187686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,1024,0.06662222411897448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,256,0.009277332988050249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,6144,0.46952268812391496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,5120,0.04210577739609612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,3072,0.02525955604182349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,6144,0.032771554258134626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,512,0.010775999890433418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,65536,0.22266843583848742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,12288,12288,0.05190044310357836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,2048,0.01132266637351778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,1024,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,10240,0.015849777393870883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,128,0.0034888887570963967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,2048,0.03741688860787286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,3584,0.13435199525621203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,2048,0.04090133309364319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,32,0.002704000100493431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,1536,0.020260444945759244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,512,51200,0.28556977377997506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,2560,0.056636446052127414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2048,65536,0.6631377538045248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,5120,0.017988445030318368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,32,0.0029813332690132987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,3072,0.12951821751064724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,3584,0.013453333742088742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,8192,0.22373334566752115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,1024,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,65536,0.06963377528720431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,256,0.012317332956526013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,4096,0.013839110732078552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,12288,0.009401777552233802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,256,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,8192,0.016399999459584553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,1024,0.0037573331760035623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,51200,0.1924746698803372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,8192,0.19995200634002686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,8192,0.0906897783279419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,65536,0.33223999871148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,6144,0.05604622099134657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,4096,0.03323288758595785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,65536,0.06965688864390056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,51200,0.05316177672810025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,256,0.006842666616042455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,1536,0.15253155761294895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,12288,0.04550933175616794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,1024,0.006723555425802867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,256,0.10400888654920791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,10240,0.16072621610429552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,12288,2.2091039021809897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,7168,0.018058665924602084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,1536,0.10568888982137044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,128,0.006494222415818109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,2048,0.014215999179416232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,4096,0.008567999634477828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,16384,0.05310844381650289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,2560,0.030681778987248737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,5120,0.00758844448460473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,6144,0.02882044514020284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,32,128,0.002630222174856398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,3072,0.5809475580851237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,512,0.00517155561182234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,10240,0.020779555042584736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,12288,0.5888471073574489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,4096,0.020703110429975722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,3072,0.006701333241330252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,7168,0.18558222717709014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,8192,0.889201800028483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,3072,0.008273777862389883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,65536,0.7706666522555881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,1536,0.3535457717047797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,3072,1.005400021870931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,64,0.006703111032644908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,65536,0.5968071089850532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,128,0.1974782281451755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,65536,0.6719262335035535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,768,0.005882666756709416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,65536,6.306666480170356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,51200,23.662322998046875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,10240,0.062302225165896945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,256,0.005514666852023866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,12288,0.48369598388671875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,256,0.032002667586008705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,4096,0.18996800316704643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,8192,0.04976444443066915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,1536,0.05166222320662605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,3584,0.00959288908375634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,256,0.003662222375472387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,3584,0.012446222205956778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,5120,0.006760889043410619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,8192,0.012166221936543783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,4096,0.03978311022122701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,256,0.1128657791349623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,4096,1.1948728561401367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,51200,16384,11.635771857367622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,6144,0.013070222404268054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,128,0.0047102222839991255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,12288,0.0526702238453759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,6144,0.01883555617597368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,4096,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,16384,0.10114310847388373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,10240,0.08417155345280965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,5120,0.016186666157510545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,3072,0.02674755619631873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,16384,0.16604177157084146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,1024,0.01593777702914344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,768,0.00608711110221015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,3584,0.03203466534614563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,256,0.012543111211723752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,65536,0.14421155717637804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,7168,0.18483377827538383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,128,0.04943289028273689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,512,0.003551111039188173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,12288,0.018265777164035372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,16384,0.3552195496029324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,10240,0.018937778141763475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,12288,0.3926488823360867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,1536,0.006050666587220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,6144,6144,0.017154667112562392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,51200,0.09081155723995632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,768,0.009646221995353699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,4096,0.11427111095852321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,2560,0.11586666107177734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,768,0.05468533436457316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,768,0.014418666561444601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,5120,0.010151999692122141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,64,0.005595555735958948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,7168,0.017888888716697693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,10240,0.2473048898908827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,3072,0.2706604533725315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,16384,0.02667466633849674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,256,0.011566222541862063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,2560,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,128,0.00424177779091729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,5120,0.010248888697889116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,51200,0.7577244440714518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,2048,0.03347555465168423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,2048,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,6144,0.0885475542810228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,10240,0.02664088871743944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,1536,0.012527999778588613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,4096,0.3615057733323839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,768,0.5966489050123427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,768,0.00646666685740153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,1024,0.019521777828534443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,6144,0.02423911127779219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,10240,0.05007911059591505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,768,0.004287111262480418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,6144,0.34049778514438206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,1024,0.0069653333889113525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,768,0.024731554918819006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,5120,0.0994488861825731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,4096,0.07690222395790948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,256,0.003383111208677292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,4096,0.1875457763671875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,3072,0.10220533609390259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,512,0.005947555518812603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,512,0.031230221192042034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,2048,0.008181333541870117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,51200,0.015779554843902588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,768,0.08942577573988174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,12288,0.06483288606007893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,2048,0.4749129083421495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,64,0.01297511077589459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,32,0.007641777396202087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,16384,0.0556817783249749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,512,0.006435555716355641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,8192,0.08911999728944565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,512,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,16384,0.2517928812238905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,6144,0.01018666641579734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,65536,5.004357231987847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,3072,0.02760800056987339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,10240,0.024881778491867915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,7168,0.08909066518147786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,256,0.004084444294373195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,5120,0.02424888809521993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,768,0.0061928890645504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,768,0.006705777926577463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,5120,0.06167822413974338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,4096,0.026072000463803608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,7168,0.10904355843861897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,8192,0.025070221887694463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,128,0.00435288871328036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,16384,0.07427910963694255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,256,0.0042275554604000514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,3072,0.2014862166510688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,1536,0.006416000011894438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,12288,1.1011750962999132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,2560,0.0696622199482388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,2048,0.04874755607710945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,2048,0.009017777939637503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,16384,0.6783271365695529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,1024,0.34548266728719074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,2048,0.008050666915045844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,64,0.005153777698675792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,512,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,7168,0.0646302236451043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,65536,0.48621511459350586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,2560,0.007012444237867991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,65536,0.10903111431333755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,3584,0.03038933210902744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,64,0.0026746667507621977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,7168,5.133647918701172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,128,0.003620444486538569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,1024,0.005151999907361137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,10240,0.14132711622450086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,768,0.02349066734313965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,768,0.008245333201355403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,3584,0.025846223036448162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,12288,0.4046764373779297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,2048,0.028204444381925795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,4096,0.009529777699046666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,10240,0.024223110742039148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,51200,0.3200026618109809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,3584,0.00980444418059455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,8192,0.07635821898778279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,64,0.002990222225586573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,6144,0.019840000404251944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,7168,0.3095164563920763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,64,0.00574933323595259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,256,0.03775733378198411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,5120,0.15458044740888807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,12288,0.5469644334581163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,768,0.009000000026490953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,6144,0.0355662206808726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,64,0.01183733344078064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,2560,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,2048,0.016640888320075143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,512,0.03849244448873732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,65536,0.07591022385491265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,512,0.004331555631425646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,128,0.004573333180612988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,5120,0.020984888076782227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,51200,0.05321777860323588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,51200,0.20229244232177734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,3584,0.20095466242896187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,10240,0.08952177895439996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,256,0.007000000112586551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,12288,0.02774133284886678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,16384,0.025832888152864244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,7168,0.02169688873820835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,51200,0.05155644483036465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,3584,0.008624888956546783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,8192,0.7374222013685438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,128,0.011215111447705163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,128,0.010947555303573608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,4096,0.42186933093600804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,3072,0.03758666581577725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,64,0.01720888912677765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,3072,0.029859556092156306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,256,0.005459555735190709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,32,0.03015822172164917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,6144,3.908362494574653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,256,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,16384,0.06678844160503812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,2560,0.05543733305401272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,3072,0.1365333398183187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,1536,0.026405332816971674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,16384,0.03354933195643955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,3584,0.49264801873101133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,6144,0.01090311093462838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,6144,0.048231111632453076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,6144,0.01734133395883772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,12288,0.29753510157267254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,10240,0.02588711182276408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,2048,0.006172444257471297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,12288,0.08797066741519505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,51200,0.05440355671776665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,3072,0.010114666488435533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,64,0.008253333469231924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,512,0.003415999934077263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,1536,0.006238222122192383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,6144,0.04660355713632372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,12288,0.006020444548792309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,12288,0.09141066339280869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,6144,0.01087288889620039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,1536,0.010486221975750394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,512,0.005260444349712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,5120,0.007599111232492659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,5120,0.05497866537835863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,65536,0.13134222560458714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,4096,0.01899111105336083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,512,0.0037760001917680106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,2560,0.007459555235173967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,64,0.059847997294531934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,16384,0.5375617874993218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,256,0.04791733291414049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,128,0.0028702221396896574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,768,0.0177484436167611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,65536,0.033752000994152494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,4096,0.19641510645548502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,1536,0.00721155520942476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,51200,0.0961022244559394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,65536,0.1480791171391805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,768,0.007633777956167857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,10240,0.02132355504565769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,65536,0.16610222392612034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,128,0.0027342221389214196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,7168,0.13982844352722168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,12288,0.08396355311075847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,65536,0.07825511031680636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,12288,0.430785788430108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,256,0.0029813332690132987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,10240,0.023187556200557288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,12288,0.30119021733601886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,768,0.5796435674031576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,1536,0.015547555353906421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,6144,0.045250667466057666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,12288,0.019689778486887615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,4096,0.17612000306447348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,3072,0.013644444445768992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,8192,0.027327110370000202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,7168,0.7095822228325738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,6144,0.5868310928344727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,256,0.007311111523045435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,512,0.005970666805903117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,10240,0.013533333109484779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,768,0.027591110931502447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,2048,0.033448888195885554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,10240,0.014223110344674854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,7168,2.097010718451606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,3072,0.04164800047874451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,2560,0.019151111443837483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,4096,0.00997155573632982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,5120,0.25736623340182835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,256,0.01181244436237547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,6144,0.19824711481730142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,3072,0.28399909867180717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,7168,1.3412728839450414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,65536,0.2958515485127767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,3072,0.14865688482920328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,6144,0.12003110514746772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,12288,0.047489778863059155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,32,0.006894222150246303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,1536,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,5120,0.012456888953844706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,16384,0.03492000036769443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,8192,0.07113422287835015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,512,0.003591999825504091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,5120,0.02283199959331089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,64,0.0030453333424197305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,64,0.004950222041871813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,12288,0.2596835560268826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,7168,0.01757600075668759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,51200,0.1352657741970486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,256,0.021673777037196692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,65536,0.14163466294606528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,10240,0.2636142306857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,128,0.003028444531891081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,51200,1.2500239478217232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,3584,0.008754666480753157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,10240,0.04197777642144097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,7168,0.18664799796210396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,7168,0.05559911330540975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,12288,0.0673324465751648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,2048,0.014900444282425774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,1024,0.011030221978823343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,5120,0.013944889108339945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,7168,0.21808355384402803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,65536,0.9147706561618381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,51200,0.1929386721716987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,12288,0.021728888154029846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,7168,0.04978133241335551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,32,0.03234311276011997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,16384,0.6084639761182997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,3072,0.008824889030721452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,3584,0.06555466519461738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,4096,0.024688889582951862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,16384,0.025839999318122864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,16384,0.09848888715108235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,10240,0.647464911142985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,12288,0.028246220615175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,7168,2.447745853000217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,2560,0.026484444737434387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,32,0.003232888877391815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,5120,0.15684444374508327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,1536,0.6784826914469401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,5120,0.013567999833159976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,2560,0.4561591148376465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,7168,0.012191111014948951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,51200,0.025411556164423626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,7168,0.15055999490949842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,6144,0.8762942420111762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,1536,0.006025777922736273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,768,0.006786666810512543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,1536,0.035213334692849055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,128,0.005888000130653381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,2560,0.013620444469981723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,7168,0.012814222110642327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,3072,0.04172711239920723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,16384,1.4012230767144098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,12288,0.02500088844034407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,768,0.02409244411521488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,51200,1.3363297780354817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,4096,0.010379555324713389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,512,0.003568888952334722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,2560,0.5028746922810873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,64,0.02693422304259406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,5120,0.04059200154410468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,128,0.029183109601338703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,3584,0.0177484436167611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,32,0.011454222102959951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,3072,0.06626933150821261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,65536,1.4063235388861761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,2560,0.008954666554927826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,64,0.003806222230195999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,1536,0.042583111259672374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,32,0.052807112534840904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,512,0.006320888797442119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,6144,0.04444800151718987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,3072,0.010037333601050908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,256,0.007807999849319458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,4096,0.017772444420390658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,3072,0.010656000011497073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,10240,0.17827467123667398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,10240,0.032334221733940974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,32,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,7168,0.01143022212717268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,7168,0.04495822058783638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,10240,0.016578666037983365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,1536,0.006411555740568373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,1024,0.0053093334039052325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,7168,0.10629510879516602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,5120,0.01440088947614034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,8192,0.0853288902176751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,5120,0.11268177297380234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,16384,0.1821999947230021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,65536,2.8023164537217884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,5120,12288,0.36632800102233887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,7168,0.01865600049495697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,128,0.14481511380937365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,32,0.009733333355850643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,10240,0.08228800031873915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,3072,0.027828445037206013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,64,0.00869955536392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,1536,0.0063680000603199005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,256,0.025610667135980394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,4096,0.19156266583336723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,8192,0.11814399560292561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,3072,0.0074471111098925276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,3584,0.2319120036231147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,64,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,2048,0.007227555745177799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,8192,0.03788888785574172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,1536,0.012936000194814471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,1024,0.004088888979620404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,768,0.008775110873911116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,16384,0.3139306704203288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,64,0.005612444546487596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,3584,0.04610933197869194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,8192,2.942966249254015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,10240,0.11026133431328668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,16384,8.994478861490885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,5120,0.21734044286939833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,10240,0.5571244557698568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,10240,0.04913777775234646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,3072,0.032560000816980995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,8192,0.009122666385438707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,16384,1.6025217903984919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,12288,1.000790172153049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,51200,1.1427893108791776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,2560,0.010431110858917236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,6144,0.24203732278611925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,12288,3.7287669711642795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,8192,0.023947555157873366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,7168,0.008570666942331526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,1024,0.014296889305114746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,2048,0.011121778024567498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,8192,0.07519555754131742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,32,0.007625777688291337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,32,0.012114666402339935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,32,0.10405066940519546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,10240,0.015189333094490899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,10240,0.22032799985673693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,128,0.015264888604482016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,2560,0.12186222606235081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,7168,0.08752889103359646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,3584,0.47230487399631077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,10240,0.03801066676775614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,12288,0.20177688863542345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,7168,0.014300444059901767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,1024,0.012085333466529846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,10240,0.04340622160169813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,16384,0.04823644293679131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,3072,0.09144000212351482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,2560,0.007237333390447829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,768,0.005606222069925732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,6144,0.027432888746261597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,3584,0.014330666926172046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,51200,0.21923466523488364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,2560,0.006262222097979651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,256,0.005163555757866966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,3584,0.3402400016784668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,51200,0.6297866503397623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,512,0.01980888843536377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,3072,0.03400355577468872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,32,0.005027555757098728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,1536,0.04622488882806566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,512,0.053908444113201566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,32,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,65536,0.03497599893146091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,2560,8.348960876464844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,2048,0.007174222005738153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,32,0.003245333416594399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1024,51200,0.5657857788933648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,256,128,0.0047280001971456735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,768,0.012868444124857584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,6144,0.0543431109852261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,256,0.00906311141120063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,1536,0.046262221203909985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,768,0.020120888948440552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,256,0.0031973332580592898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,7168,3.5634549458821616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,512,0.01626844373014238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,2048,0.024689777029885188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,6144,0.20371554957495797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,10240,0.006692444284756978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,4096,0.02181333303451538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,4096,0.022265776991844177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,5120,0.14263644483354357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,1024,0.043486220969094165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,8192,0.02134399943881565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,5120,0.9880818261040581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,4096,0.0466542210843828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,2048,0.008610666626029545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,51200,134.36416286892361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,5120,0.7054542435540093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,256,0.004119111018048393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,128,0.004798222333192825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1024,8192,0.011367110742463006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,7168,0.020210666788948905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,2560,0.4897751278347439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,65536,2.619850582546658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,12288,0.31669243176778156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,256,0.0034560000316964257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,8192,0.01648266613483429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,768,0.01740977830357022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,2560,0.08368177546395196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,65536,0.8681848843892416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,1024,0.005593777944644292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,1024,0.005700444595681296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,64,0.005963555640644497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,16384,0.2847191227806939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,8192,0.02753955622514089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,3072,0.23038223054673937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,5120,0.02327466673321194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,5120,0.40089599291483563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,4096,0.010482666393121084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,12288,0.2694888909657796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,768,0.008773333496517604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,2560,0.00775911079512702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,256,0.015343111422326831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,2560,0.007165333463086023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,3072,0.01603911154799991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,7168,0.02922933300336202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,12288,0.1778764459821913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,12288,0.04541866646872627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,2560,0.025901334153281316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,128,0.015961777832772996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,128,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,12288,0.032398223876953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,2560,0.012219555675983429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,5120,0.017488888568348356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,3584,2.145696004231771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,3072,0.040390223264694214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,128,0.02313066687848833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,4096,0.014653333359294467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,768,0.22001066472795275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,128,0.006992888947327931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,5120,0.011690666278203329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,51200,0.1275955571068658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,64,0.016211556063758004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,128,0.004510222209824456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,512,0.01667288939158122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,64,0.002569777684079276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,12288,1.8850729200575087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,8192,0.060912887255350746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,64,0.0033537778589460584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,3584,0.025916443930731878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,7168,0.3657635582817926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,1536,0.006084444622198741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,5120,0.24317333433363172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,768,0.019742222295867074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,5120,0.17227644390530059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,2560,0.0069573331210348345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,64,0.008821333448092142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,2560,0.08725511365466648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,65536,2.6025013393825955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,2048,0.01995377739270528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,3072,0.22403022978040907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,8192,0.016585777203241985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,128,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,2560,0.04808977908558316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,65536,0.3980613284640842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,256,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,51200,0.06325599882337782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,16384,0.08203111092249553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,5120,0.049343112442228526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,1024,0.03011022342575921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,7168,0.319123559527927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,256,0.017879999346203275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,6144,0.2710079881880018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,5120,0.05920444594489204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,512,0.07328533464007907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,2048,0.04598311252064175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,3584,0.0243520008193122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,512,0.06536444690492418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,4096,2.5817457834879556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,3072,0.010176888770527309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,12288,0.03835466504096985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,256,0.03612444467014737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,6144,0.25982220967610675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,256,0.004744889007674323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,4096,0.006753777878151999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,16384,0.28335287835862905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,65536,0.08161244127485487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,1536,0.006738666858938005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,1536,0.005114666703674528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,3584,0.09948177470101251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,1536,0.024358222881952923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,256,0.0029484445436133277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,5120,0.056001775794559054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,2560,0.012786666552225748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,3072,0.02125066684352027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,256,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,1024,0.009789333575301701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,2048,0.010067555639478896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,64,0.006619555668698416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,7168,0.020113777783181932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,8192,0.1536088916990492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,768,0.022843556271659002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,6144,0.017243555850452848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,6144,0.2665884494781494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,2048,0.010321777727868822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,10240,1.7370604409111872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,3584,0.014634667171372307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,768,0.04615199897024366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,256,0.00830844458606508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,512,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,8192,0.008275555239783393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,6144,7.8362986246744795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,6144,0.016890666551060148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,16384,0.1657279994752672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,8192,0.029211554262373183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,32,0.0057839999596277876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,64,0.0038124442928367188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,8192,0.03319911162058512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,32,0.00555377784702513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,8192,0.014127999544143677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,7168,0.02407199972205692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,1536,0.014159111513031853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,1024,0.02612888813018799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,8192,0.20570845074123809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,7168,0.032913777563307024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,256,0.0041511112617121804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,768,0.022533333963818018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,1024,0.008186666501892937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,1024,0.008526222573386298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,16384,0.08831377824147542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,8192,0.05968533621893989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,1024,0.005331555588377847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,12288,0.01017155581050449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,1536,0.009097778134875828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,1536,0.019761777586407132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,5120,0.011183111204041375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,8192,0.23870489332411024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,3072,0.026562665899594624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,7168,0.06307644314236112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3584,2560,0.009911110831631554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,256,64,0.0026062221990691293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,1024,0.010635555618339114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,3584,0.024197333388858374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,3072,0.05236977669927809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,6144,0.04106222258673774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,2048,0.014145778285132514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,128,0.004571555389298333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,4096,0.020788444413079154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,6144,0.013937777943081327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,8192,0.039261334472232394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,256,0.007357333269384172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,4096,0.03478844298256768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,5120,0.018197332819302876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,7168,0.02273955610063341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,3584,0.00980533328321245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,65536,0.19543022579616973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,4096,0.022104000051816303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,512,0.007893333004580604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,6144,0.018647111124462552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,3584,0.061760889159308545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,128,0.00508266646001074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,12288,0.01819022165404426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,3072,0.017467555072572496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,65536,0.18880089124043783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,3072,0.010076444182131026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,64,0.007627555893527136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,768,0.007682667010360294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,12288,8.184154934353298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,51200,14.480321248372396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,8192,7.548954433865017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,2560,0.019325332509146798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,64,0.0042408886882993905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,4096,0.02977688776122199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,256,0.005728000154097875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,1536,0.018976888722843595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,4096,0.14483643902672663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,2560,0.05240622162818909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,4096,0.01758666667673323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,32,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,12288,0.028406222661336262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,51200,0.46413776609632706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,10240,0.3387048774295383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,1024,0.007270221908887227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,51200,0.7678577635023328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,2048,0.006569777925809224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,7168,0.04031377699640062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,8192,2.901403639051649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,16384,0.02680711117055681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,65536,0.33309245109558105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,3072,0.03558489018016391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,1536,0.09328711032867432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,3072,0.1234640015496148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,65536,0.08655466636021932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,8192,0.07099911239412096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,65536,1.7532044516669378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,6144,0.015322667029168872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,32,0.004631999880075455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,12288,1.2361893124050565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,51200,0.11868711312611897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,1536,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,4096,0.10494666629367405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,4096,0.015432889262835184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,16384,0.023756444454193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,12288,0.5437368816799587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,768,0.02019733356104957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,65536,1.2644355561998155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,64,0.0074346669846110875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,10240,0.016531555189026725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,1024,0.010312889185216691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,32,0.003047111133734385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,2048,0.02424533334043291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,768,0.018868444694413077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,51200,0.14391467306349012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,5120,0.0158897770775689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,768,0.038154668278164335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,3072,0.011345778074529437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,1536,0.02404088940885332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,512,0.005581333405441708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,128,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,4096,0.008032889001899296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,3584,0.28098487854003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,512,0.013982221484184265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,4096,0.19759821891784668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,3072,0.09352978070576985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,3584,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,3072,0.024785778588718835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,51200,1.229532453748915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,51200,0.07101333141326904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,2048,0.02382755610677931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,16384,0.012797333300113678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,3072,0.006893333461549547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,256,0.011565333439244164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,7168,0.02180888917711046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,3584,0.02067733307679494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,10240,0.58642578125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,51200,0.3464115460713704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,65536,0.08089955647786458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,12288,0.15594755278693304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,51200,0.05769866704940796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,10240,0.10142578019036187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,4096,0.04721777637799581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,4096,0.418271991941664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,3072,0.03837955660290188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,7168,0.03675466775894165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,32,0.010965333216720156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,2560,0.006698666761318843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,7168,0.02389777700106303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,32,0.01800622211562263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,1024,0.01186222251918581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,256,0.00518222235971027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,5120,0.028861333926518757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,4096,0.013397333522637686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,3584,0.4454142252604167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,512,0.15319289101494685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,2048,0.009582222335868413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,10240,0.1257271104388767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,10240,0.016711110870043438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,6144,0.8112311363220215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,51200,0.06915644142362806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,7168,0.06948977708816528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,4096,0.2545608944363064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,6144,1024,0.008024888734022776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,512,0.1650951173570421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,10240,0.03176711002985636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,3584,0.01425066590309143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,3072,0.013185777597957186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,12288,0.16917778386010063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,3584,0.011237333218256632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,3584,0.13286399841308594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,12288,0.6492897669474283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,5120,0.012791111237472959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,2560,0.09467289182874893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,7168,0.013025778035322825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,2048,0.7568426662021213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,256,0.009955555200576782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,2560,0.0577404432826572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,8192,0.08859378099441528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,1536,0.012092444631788464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,2048,0.00847466703918245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,128,0.010130667024188572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,1536,0.04781422350141737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,51200,0.04351466563012865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,1024,0.03133866522047255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,7168,0.016235555211702984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,12288,0.6030000050862631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,4096,0.39744088384840226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,4096,0.010136889086829292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,768,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,64,0.003026666740576426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,768,0.039926220973332725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,64,0.005643555687533484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,64,0.025546666648652818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,64,0.00776800016562144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,12288,0.02033333314789666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,2560,0.02474400069978502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,512,0.01092177795039283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,8192,0.17328533861372206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,5120,0.3522026538848877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,2560,0.010393777655230628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,3584,0.06973244746526082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,3584,0.04233422213130527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,6144,0.05304888884226481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,64,0.00563733321097162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,32,0.011771555576059552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,5120,8192,0.28063644303215873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,3584,0.008359111017651027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,2560,0.5433075692918565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,1536,0.004729777574539185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,3584,0.027121777335802715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,2560,0.11801777945624457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,3584,0.03927466604444716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,512,0.02053777707947625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,768,0.013639999760521783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,16384,0.06617866622077094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,7168,0.015523556205961438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,3584,0.018564444449212816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,3072,0.9142835405137805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,5120,0.014326221413082547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,12288,0.23977687623765734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,512,0.0036844445599450003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,65536,3.411567052205404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,10240,0.42560709847344297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,2560,0.024315555890401203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,65536,0.3171653217739529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,5120,0.01441511180665758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,4096,0.2073048883014255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,51200,2.5057254367404513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,512,0.04000000158945719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,5120,0.054094221856859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,256,0.008479110896587372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,512,0.008450667063395182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,768,0.006683555742104848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,128,0.004868444469239977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,128,0.008385777473449707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,51200,0.06969244612587823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,4096,0.009043555292818282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,2560,0.00832266691658232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,2048,0.015984000431166757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,32,0.0058551111982928375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,12288,0.02514222264289856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,768,0.007632888853549957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,32,0.0032888888898822996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,2048,0.01642311116059621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,4096,0.0173333336909612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,7168,0.03640444411171807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,8192,0.026943999860021803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,5120,0.4339599874284532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,512,0.15851022137535944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,768,0.2594035466512044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,1536,0.18261333306630453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,6144,2.8394987318250866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,10240,0.01945333348380195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,51200,20.572674221462673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,7168,0.014749333262443542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,3584,0.009544889132181803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,6144,0.2665964497460259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,10240,0.028060442871517603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,64,0.0029031110720502008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,65536,0.07235466771655612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,512,0.017085333665211994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,12288,5.578403472900391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,1536,0.007167999943097432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,768,0.016484444340070088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,5120,0.06775644752714369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,16384,0.3950968848334418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,65536,0.2905777825249566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2560,768,0.02036977807680766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,7168,0.3589297665490045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,10240,0.00814844419558843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,2560,0.19543377558390299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,1536,0.010151111417346531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,4096,0.007279111279381647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,65536,0.0685102211104499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,768,0.3883333206176758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,2048,0.00776088900036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,2048,0.007032000356250339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,5120,0.009588444398509132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,768,0.003721777763631609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,64,0.06247199906243218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,5120,0.12464266353183323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,1536,0.06792266501320733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,3072,0.01440533333354526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,64,0.011559111376603445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,2560,0.07364533344904582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,512,0.013374221821626028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,2048,0.014662222729788886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,7168,0.006706666615274217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,51200,0.7213902473449707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,3584,0.04420800010363261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,51200,16.553580390082466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,51200,1.450548489888509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,32,64,0.003506666670242945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,51200,0.04657155606481764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,65536,0.295015123155382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,6144,0.05635199944178263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,1024,0.006441777778996362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,2048,0.02253155575858222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,12288,0.06191200017929077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,2560,0.009027555584907532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,5120,0.017108443710539076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,16384,0.02148888839615716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,7168,0.40757155418395996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,768,0.01314133322901196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,1536,0.013049778011110095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,5120,1.7808596293131511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,10240,1.0465786192152235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,256,0.0038702223036024305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,4096,0.03660444418589274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,768,0.006443555570311016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,4096,0.010680889089902243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,65536,0.08637422323226929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,16384,0.07319377528296576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,512,0.0337137778600057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,5120,2.3416355980767145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,8192,0.33796710438198513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,65536,0.19347111384073892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,8192,2.826375961303711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,3072,0.01256622208489312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,256,0.03213511241806878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,32,0.0033039999090962936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,7168,0.23541333940294054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,10240,0.022760000493791368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,1536,0.05516977773772346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,10240,0.05615199936760796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,6144,0.16449688540564641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,65536,0.5807564523484972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,2048,0.007107555866241455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,1536,0.020840888222058613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,2560,0.0074275558193524676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,5120,0.17807377709282768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,7168,0.01183199965291553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,7168,0.010796444283591377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,4096,0.07657688856124878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,1024,0.011615999870830111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,512,0.00904088881280687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,2560,0.011886222494973077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,256,0.1342888938056098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,1024,0.01202933324707879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,16384,0.017708443933063082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,6144,0.12210044595930312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,1536,0.011828444070286222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,3584,0.054775999652014844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,768,0.01535999940501319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,3584,0.028093334701326158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,10240,0.052629331747690834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,768,0.06670755810207792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,10240,0.028051555156707764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,8192,0.04735022121005588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,8192,0.014458666245142618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,12288,0.6156586541069878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,65536,7168,0.24346754286024305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,16384,0.0365937782658471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,3072,0.04033777780003018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,3584,0.007610666255156199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,5120,0.03375022278891669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,1024,0.030192000998391047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,7168,0.14427467187245688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,2048,0.013292444248994192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,512,0.006513777706358168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,10240,0.014386667145623101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,1024,0.00666133314371109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,51200,0.21760088867611357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,3584,0.2182391087214152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,3072,0.008808000220192803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,512,0.005392888767851724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,2048,0.005746666755941179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,512,0.0033599999215867785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,7168,0.01977066695690155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,1024,0.005392888767851724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,8192,0.12385155094994439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,512,0.005967999911970562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,8192,0.2652195559607612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,2048,0.009836444424258338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,32,0.11724266741010879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,512,0.00554933316177792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,1536,0.014635556273990206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,3072,0.018823999497625563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,1536,0.010457778142558204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,2560,0.010144000252087912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,10240,0.06787288851208158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,1536,0.11199466387430827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,768,0.005828444328573015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,65536,0.34468711747063535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,4096,0.016140444411171805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,768,0.027581334114074707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,512,64,0.0026515554636716843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,10240,0.016530666086408827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,3072,0.02667733364635044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,32,0.005015111217896144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,256,0.013926222920417786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,12288,0.16094222333696154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,512,0.0032755554550223877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,16384,0.07421600156360202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,3584,0.05179466803868612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,768,0.00519466648499171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,65536,2.4618070390489364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,7168,0.01609866652223799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,51200,342.5164388020833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,2560,1.0106230841742623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,3584,1.0604267120361328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,256,0.005522666705979242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,2560,3.168384975857205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,128,0.03434933225313822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,1536,0.1521093315548367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,768,0.004281777888536453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,3072,0.05702310800552368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,2048,0.008917333351241218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,16384,0.12632889217800564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,256,0.004378666480382283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,1024,0.025496888491842482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,51200,0.23655290073818633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,3072,0.013068444199032254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,3072,0.07739110787709554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,51200,0.14224089516533747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,1536,0.5838364495171441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,51200,0.7632124688890246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,768,0.004847110973464118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,65536,0.7168862024943033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,7168,0.04157955447832743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,2048,0.0808968875143263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,1536,0.008991999758614434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,7168,0.08954488568835789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,4096,0.037616888682047524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,128,0.00498933345079422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,2560,0.005568000177542369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,1024,0.006561777657932705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,7168,32,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,6144,1.9098924001057942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,512,0.1749253273010254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,256,0.0032213332338465583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,8192,0.2086177799436781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,10240,0.16282044516669378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,5120,0.3306364483303494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,128,0.0075235557225015424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,256,0.003922666526503033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,16384,0.03977244430118137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,4096,0.24457687801784941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,64,0.013024888932704926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,5120,0.04191911220550537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,3072,0.02167466613981459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,51200,0.09225689040289985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,128,0.21342666943868002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,128,0.013882666826248169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,12288,0.2761190997229682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,10240,0.023734221855799358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,2048,0.01051288925939136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,2048,0.06146577994028727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,768,0.04541600081655714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,8192,0.012445333103338877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,65536,0.2952435546451145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,1024,0.015119999647140503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,8192,0.18770578172471789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,3072,0.009914666414260864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,2048,0.006308444258239534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,5120,0.5261591275533041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,8192,0.01365244471364551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,2560,0.04142755601141188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,3072,0.07341066996256511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,768,0.005560889012283749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,768,0.01182933317290412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,128,0.1479697757297092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,4096,0.011036444041464063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,128,0.06788444519042969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,10240,0.2884257899390327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,1536,0.030180444320042927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,4096,0.007115555306275685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,12288,0.03387377659479777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,51200,0.5261759757995605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,2560,0.009148444566461775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,512,0.014525334040323893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,5120,0.009578666753239101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,16384,0.5416657659742568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,768,0.01665866706106398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,12288,0.026144888665941026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,64,0.0027786667148272195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,256,0.003752888904677497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,32,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,2048,0.0058044443527857465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,10240,0.07256088654200236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,128,0.0070266665683852295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,512,0.009400889277458191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,3584,0.011606222225560082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,256,0.021525333325068157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,7168,0.2471617857615153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,12288,0.017872000734011333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,64,0.005954666684071223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,3584,0.016781333420011733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,1024,0.10723911391364203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,1536,0.15690667099422878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,16384,0.02251822253068288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,5120,0.010210666391584609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,16384,0.12404355737898086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,8192,0.27969421280754936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,3584,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,5120,0.15123021602630615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,4096,0.00798044436507755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,1536,0.6004186736212836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,1024,0.03962399893336826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,3584,0.42799467510647243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,128,0.0046693334976832075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,12288,0.011195555329322815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,768,0.0058551111982928375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,65536,3.3845698038736978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,65536,0.14542399512396917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,12288,0.05293066634072197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,6144,0.6623297797309028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,51200,0.18569689326816133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,65536,2.149553722805447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,65536,0.08811910947163899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,512,0.01811644434928894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,3072,0.009489778015348647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,16384,0.04010488920741611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,1024,0.011854222251309289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,2048,0.017543110582563613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,10240,0.17832088470458984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,10240,0.2427875465816922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,1024,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,5120,0.018086221483018663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,3072,0.0087333329849773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,32,0.023345778385798138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,3584,0.07106310791439481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,512,0.0400622222158644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,51200,5120,1.392075538635254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,1536,0.007321777443091075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,64,0.0070568886068132185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,3072,0.025157334076033697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,3072,0.1242613395055135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,12288,0.0753671129544576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,32,0.005260444349712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,256,0.007129777636792924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,5120,0.024472888973024156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,1024,0.00600355573826366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,768,0.008602666358153025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,128,0.006212444355090459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,2560,0.05822755230797661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,4096,0.016846223009957206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,5120,0.2243759897020128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,1536,0.005578666511509154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,3072,0.015666666958067153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,32,0.0029537777105967202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,1536,0.00830488900343577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,1536,0.004785777793990241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,65536,0.08149688773685031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,3584,0.024990222520298425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,1536,0.009177777502271865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,256,0.016992888516849942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,768,0.003957333250178231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,128,0.004045333299371931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,7168,0.005991999887757831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,10240,0.023364444573720295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,8192,0.050535112619400024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,1536,0.00997866690158844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,512,0.005478222337034013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,256,0.004938666605287128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,3072,0.06344799862967597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,1536,0.021309332715140447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,3584,0.040792889065212674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,7168,0.013528889252079858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,768,0.01352266636159685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,16384,0.07579021983676486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,1024,0.0037724446091387006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,32,0.008192000289758047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,2048,0.4529751141866048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,256,0.009081777599122789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,6144,0.018698667486508686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,2560,0.009101333717505137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,10240,0.19414399729834664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,128,0.003527111063400904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,8192,0.014787554740905762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,51200,2.607751210530599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,1536,0.007225777539942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,12288,3.612415313720703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,2560,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,128,0.005761777775155173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,16384,0.04150222076310052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,128,0.00282222218811512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,7168,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,768,0.02700000007947286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,6144,0.03879377908176846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,768,0.0058711109062035876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,3072,0.03238666719860501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,16384,0.2247377766503228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,12288,0.040454222096337214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,3072,0.022601778308550518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,51200,1.2321546342637804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,128,0.00851288851764467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,10240,0.23807999822828504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,8192,0.01887022289964888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,32,0.006752000086837345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,512,0.0074720001882976955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,64,0.002995555599530538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,2560,0.8117538028293185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,51200,0.32714933819240993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,7168,0.017490666773584154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,2560,0.009486222432719337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,8192,0.6630951033698188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,1536,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,6144,0.01331555595000585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,512,0.21402400069766572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,3584,0.060545775625440806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,512,0.006793777975771163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,3072,0.007155555817815993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,3584,0.01182755579551061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,2048,0.00602311102880372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,512,0.04461955693033007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,3072,0.010066666536860997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,768,0.0090524446633127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,32,0.02702755563788944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,51200,0.5030479960971409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,12288,0.02661866611904568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,10240,0.424274656507704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,512,0.027652444110976324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,1024,0.01685599982738495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,16384,0.023883556326230366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,128,0.008231110870838165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,512,0.026865777042176988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,32,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,768,0.01090311093462838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,1536,0.007348444726732042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,256,0.00565155554148886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,16384,0.021856889128684998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,2048,0.007657777931955125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,768,0.0036924444139003754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,5120,2048,0.008725333544943068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,2048,0.036200887627071805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,3072,0.03597155544492934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,256,0.01348355578051673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,4096,0.04145244426197476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,10240,0.0499288903342353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,2048,0.0075351107451650835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,51200,0.16950399345821807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,7168,0.0403733319706387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,16384,1.0353600184122722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,1024,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,64,0.00700444479783376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,12288,0.2688737710316976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,65536,0.3511813216739231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,3584,0.015280889140235053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,5120,0.024340444140964087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,512,0.005027555757098728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,5120,0.012091555529170565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,7168,0.4150071144104004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,65536,2.2541768815782333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,512,0.032983111010657415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,1536,0.005792888916201062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,64,0.006715555571847492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,3072,0.02680622206793891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,1024,0.008344888687133789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,12288,0.10614222288131714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,512,0.008777778181764815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,256,0.006722666737106111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,512,0.006047111004590988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,8192,0.03144977821244134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,6144,0.11510488722059463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,16384,0.025042666329277888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,51200,0.48415289984809023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,3584,0.0821288890308804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,7168,0.028845333390765723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,4096,0.14519288804796007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,2560,0.017488888568348356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,1536,0.007594666547245449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,512,0.07599288887447782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,4096,0.7898711098564996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,5120,0.018417777286635507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,3072,0.021164443757798936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,16384,0.059247997072007924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,128,0.0035964445107513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,1024,0.008192000289758047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,2560,0.03876533442073398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,16384,0.0648684435420566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,5120,0.15489510695139566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,8192,0.05505688985188802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,10240,0.26998400688171387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,256,0.004607111215591431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,256,0.007576888634098901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,3584,0.035014222065607704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,7168,0.015357333752844067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,512,0.0069511110583941145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,5120,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,256,0.00434755575325754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,10240,0.01722044414944119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,3584,0.021382222572962444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,65536,4.555168999565972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,3584,0.37032000223795575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,256,0.0069351109365622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,10240,0.05421511001057095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,256,0.012780444489585029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,32,0.015474667151769003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,1536,0.0063271112740039825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,8192,0.009505777723259395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,1024,0.008023999631404877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,512,0.003936888857020272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,16384,0.01205599970287747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,3072,0.18463999695248076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,8192,0.026124444272783067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,1024,0.007470221983061896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,64,0.2636471059587267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,128,0.0027795556104845474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,1024,0.008447111480765873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,5120,0.02179466684659322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,3584,0.01236533373594284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,64,0.003061333257291052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,16384,0.025080889463424683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,6144,0.015788444214397006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,3072,0.10353777805964152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,32,0.005584888988071018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,6144,0.01348088930050532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,10240,0.1033146646287706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,8192,0.03481066558096144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,64,0.005472888963090048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,10240,0.009776888622177972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,6144,0.04857688811090258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,6144,0.03007555670208401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,1024,0.08788533343209161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,7168,0.04610044426388211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,6144,0.026909333136346605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,7168,0.018974221414989896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,3584,0.053787555959489614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,2560,0.15903645091586643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,6144,0.029845333761639063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,768,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,64,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,3584,0.014357333381970724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,7168,0.018573333819707234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,4096,0.012647111382749347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,1536,0.0692088868882921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,5120,0.010625777973069085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,10240,0.023936000135209825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,128,0.005035555611054103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,512,0.027435556054115295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,128,16384,0.11444621615939671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,5120,0.03452533483505249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,7168,0.013339555925793119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,10240,2.55747201707628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,51200,0.05319377779960632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,64,0.003272888975010978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,4096,0.010648888846238455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,12288,0.01884888940387302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,1024,0.00944266633854972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,51200,0.07119555605782403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,8192,0.03132177723778619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,51200,131.03933376736111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,3072,0.017847110827763874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,32,0.0038622220357259116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,10240,0.9724595811631945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,1536,0.007409777906205919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,16384,0.030939555830425684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,7168,0.008266666697131263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,3584,0.012803555362754397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,1024,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,12288,0.017867555220921833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,1024,0.0053022222386466126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,10240,0.13547110557556152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,256,0.018408889571825664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,8192,0.07086222039328681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,7168,0.015504888362354703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,64,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,16384,0.07690222395790948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,2048,0.009947555760542551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,7168,0.02380888826317257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,51200,0.09071911043590969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,768,0.003660444584157732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,2048,0.10662400060229832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,16384,0.0741528868675232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,768,2.753298653496636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,3072,0.03548089000913832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,1536,0.03657866517702738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,1024,0.023334221707450017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,2048,0.06775999731487699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,6144,0.025084444218211707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,256,0.00868266655339135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,256,0.0040960001448790235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,2048,0.011965333587593503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,7168,0.022058667408095464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,1024,0.005930666708283954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,5120,0.14990844991472033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,2048,0.014726221561431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,10240,0.12324978245629205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,65536,0.3446311155954997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,256,0.007299555672539606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,12288,0.22832889027065703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,128,0.0028035555862718155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,1536,0.008172444171375698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,51200,0.1993244489034017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,1536,0.0749902195400662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,12288,0.03090488910675049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,8192,0.10649510886934067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,7168,0.21667199664645723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,128,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,8192,0.06119644641876221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,16384,0.4760382440355089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,64,0.00554933316177792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,1536,0.03425511055522495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,128,0.023404444257418316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,512,0.02083911167250739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,1536,0.005607111172543631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,51200,1.9975333743625219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,1536,0.014126222994592456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,256,0.005894222193294101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,128,0.024849777420361836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,51200,4.639311896430121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,4096,0.07296888695822822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,768,6.226871914333767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,3072,0.7767297956678602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,3584,0.010782221953074137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,16384,0.027665777338875666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,8192,0.019066666563351948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,1536,0.011271110839313932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,768,0.010657777388890585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,12288,0.054337776369518705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,51200,0.5046737988789877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,2560,0.013237333132161034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,4096,0.013581333061059317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,1024,0.007376000285148621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,128,0.014073777529928418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,1024,0.006081777728266186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,2048,0.007859555383523306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,128,128,0.0027599999060233435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,256,0.013734222286277346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,3072,0.020833777056799993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,1024,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,4096,0.00792266676823298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,64,0.005088888936572605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,6144,0.014125333891974555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,12288,0.19153155220879448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,256,0.0063493334584765965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,1024,0.0038764443662431505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,12288,0.18802577919430205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,8192,0.012108444339699216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,65536,13.252175225151909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,1024,0.06247378057903714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,16384,0.012895999683274163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,10240,0.07153777943717109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,2560,0.02583377725548214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,32,0.01442311041884952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,128,0.00619911112719112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,12288,0.00925422211488088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,8192,0.12396444214714898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,16384,0.4316942161983914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,51200,0.07079733080334134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,2560,0.012748444245921241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,12288,0.02751644452412923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,65536,0.13783289326561823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,3072,0.09150844150119358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,64,0.04584533307287428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,65536,0.12643644544813368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,65536,1.2284844716389973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,1536,0.004929777648713854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,7168,0.18850933180914986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,7168,4.055726369222005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,16384,0.01921333372592926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,2048,0.01533066729704539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,7168,0.025095999240875244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,256,0.005438222239414851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,4096,0.07452711131837633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,12288,0.10279022322760688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,768,0.007808888951937358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,768,0.12533866034613714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,6144,0.010651555326249866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,4096,0.01616533266173469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,6144,0.07764622237947252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,768,0.004923555586073133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,16384,0.043261332644356616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,512,0.18842666678958467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,1536,0.007687999970383114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,1536,0.011592888997660743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,4096,0.3632880051930745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,4096,0.008952000074916417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,3584,0.06039466460545858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,2048,0.007144889069928064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,64,0.015625778171751235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,6144,0.08176889022191365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,10240,0.2776648998260498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,65536,0.021107556091414556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,1024,0.004458666675620609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,51200,0.0590897798538208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,3584,0.009434666898515489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,16384,0.022304000125990972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,51200,0.07039999961853027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,256,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,7168,0.01107022249036365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,8192,0.06687733199861315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,4096,0.04637866550021701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,2048,0.012160888976520963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,3072,0.011542222566074796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,1024,0.07300711340374418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,2560,0.014941333068741692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,64,0.0029360000044107437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,3072,0.014689778288205465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,16384,0.021870222356584337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,6144,2.0892613728841147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,51200,0.09266933467653062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,256,0.0058124446206622655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,3072,0.0083155557513237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,16384,0.037328001525666975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,64,0.005442666510740916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,2048,0.4840986463758681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,3072,0.24515732129414877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,256,0.0036151111125946045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,2560,0.020220445262061227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,64,0.0053173332578606075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,1024,0.477891551123725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,51200,0.2016293340259128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,12288,0.08949244684643215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,1536,0.04065955678621928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,16384,0.1917804479598999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,8192,128,0.03804533349143134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,16384,0.018495111001862418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,3072,0.03130044539769491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,2048,0.00924355536699295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,3072,0.06411288844214545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,51200,8.39927503797743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,51200,1.3868818283081055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,4096,0.010796444283591377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,8192,0.04822399881150988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,16384,4.37881596883138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,16384,0.0299271113342709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,7168,0.019819556011093985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,3072,0.010942222343550788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,2560,0.015646222564909194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,3584,0.035008000002966985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,10240,0.030639999442630347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,256,0.014969777729776172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,16384,0.16397778193155924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,2560,0.007896888587209914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,3072,0.008984888593355814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,2560,0.007640889121426477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,65536,0.08963466352886623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,128,0.0074444446298811175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,51200,0.13089066081576878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,2560,0.02158311174975501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,6144,0.061182220776875816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,3072,0.018619555566045973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,3584,0.014852444330851236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,3072,0.8104746606614855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,8192,0.12007021903991699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,2048,0.0059057776298787855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,3072,0.051755554146236844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,51200,0.11223733425140381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,4096,0.03382933470937941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,65536,0.0596639977561103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,64,0.01735733283890618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,8192,0.00978311151266098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,10240,1.0521679984198677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,512,0.3517475657992893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,16384,0.11584888564215766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,10240,0.019305777218606737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,256,0.00721955547730128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,256,768,0.006538666784763336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,512,0.010797333386209277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,1536,0.006184000107977126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,3072,0.17657066716088188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,10240,0.008044444852405125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,8192,0.12331555949317084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,768,0.008640888664457533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,10240,0.07300355699327257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,768,0.023367111881573994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,512,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,7168,2.225295172797309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,51200,0.4151199923621283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,7168,0.047460443443722196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,512,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,6144,0.04401866594950358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,7168,0.11256889502207439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,3072,0.012069332930776807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,8192,0.04505777690145704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,7168,0.014127999544143677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,768,0.004335999902751711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,3584,0.008765333228641087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,2560,0.01163733336660597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,128,0.02146133283774058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,512,0.00794933322403166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,6144,0.2583884398142497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,51200,0.36073509852091473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,8192,0.18323200278811982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,7168,0.04433600107828776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,12288,0.030635555585225422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,16384,1.5985519621107314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,16384,0.2556213272942437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,8192,0.024452444579866197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,4096,0.16811022493574354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,5120,0.04448799954520332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,12288,0.011085333095656501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,3072,0.08190666966968112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,16384,0.1989742252561781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,12288,0.03100444542037116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,1024,0.009498666558000777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,768,0.03324710991647508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,64,0.006640000061856375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,51200,0.0638657808303833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,1024,0.04101955559518602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,128,0.0058133333093590206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,7168,12288,0.6979342036777072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,128,0.014979556202888489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,512,0.008539555801285638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,32,0.013191999660597907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,8192,0.01533422205183241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,7168,0.4341413444942898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,6144,0.12060177326202393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,768,0.023218666513760883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,2048,0.0069217777086628815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,8192,0.01994222237004174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,1536,0.03867111272282071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,512,0.023756444454193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,5120,0.12420533763037787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,3584,0.08038310872183906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,51200,2.8887405395507812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,51200,0.06524622440338135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3072,128,0.008388444781303406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,4096,0.034739553928375244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,128,0.002846222163902389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,65536,0.3531484338972304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,128,0.0028320000403457214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,4096,0.008226667013433244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,10240,0.4584044350518121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,16384,0.41622310214572483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,6144,0.024671999944580927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,10240,0.796870231628418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,128,0.006798222247097228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,4096,0.02852533260981242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,64,0.013468444347381592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,1536,0.007136888802051544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,128,0.0031884445084465873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,2048,0.17966577741834852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,12288,0.5236639976501465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,4096,0.008744888835483128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,1536,0.4645893308851454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,4096,0.012292444705963135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,65536,1.6977538002861872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,2048,0.09575821955998738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,65536,3.093046188354492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,128,0.06171022521124946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,16384,0.21162577470143637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,2048,0.013545778062608508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,32,0.00378577783703804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,51200,0.0537120004494985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,12288,1.192530632019043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,3584,0.02093244426780277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,1536,0.00756888919406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,1024,0.005967999911970562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,5120,0.501164436340332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,512,0.07618400123384264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,32,0.003900444342030419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,6144,0.030420445733600195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,8192,0.30868177943759495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,65536,1.3965857823689778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,10240,0.17979821893903944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,3072,0.013040000365840064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,16384,0.080677330493927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,2560,0.009709333380063375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,768,0.0037582222786214617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,768,0.0977413323190477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,16384,0.006919999917348226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,32,0.012170666621790992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,12288,0.017758222089873422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,256,0.007400000260935889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,4096,0.010263111028406356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,6144,0.06649866369035509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,8192,0.33065954844156903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,3584,0.00776622196038564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,4096,0.010811555716726514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,12288,0.08460622363620335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,256,0.006144888699054718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,5120,0.008160889148712158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,64,0.005477333234416113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,4096,0.00872622181971868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,16384,0.16620622740851507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,512,0.06651733318964641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,1024,0.025455999705526564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,1536,0.02346044447686937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,768,0.003727999826272329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,6144,0.013414222333166333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,65536,0.06197422080569797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,5120,0.030765334765116375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,6144,0.01353422221210268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,51200,0.08938222461276585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,4096,0.010120888551076254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,3584,0.014160000615649752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,64,0.00416088890698221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,3072,0.018548443913459778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,2048,0.40180622206793887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,16384,0.060460441642337374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,5120,0.057894223266177706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,6144,0.012105777859687805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,7168,0.02200266718864441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,1536,0.35005866156684023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,3584,0.012344888514942594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,4096,0.22250665558709037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,4096,0.15692355897691515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,2048,0.018926221463415358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,1536,0.010009778042634329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,7168,0.3154977692498101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,2048,0.0507262216673957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,6144,0.02889866630236308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,16384,0.006070222291681502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,32,0.04032355546951294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,256,0.004916444420814514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,3072,0.03741777936617533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,1536,0.4192710982428656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,3584,0.007651555869314406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,3072,0.00923911150958803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,768,0.007736889024575551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,256,0.0032675556010670135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,3584,0.031409776873058744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,128,0.011596444580290051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,32,0.008223999705579545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,5120,0.009549332989586724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,256,0.03641866644223531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,256,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,51200,0.05272710985607571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,768,0.008708444734414419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,51200,2.7311253017849393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,512,0.004481777962711123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,512,0.005433777968088786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2048,7168,0.04547911220126682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,32,0.007128000259399414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,6144,0.1766959958606296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,128,0.03363022208213806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,32,0.011544889046086205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,2560,0.011778666741318174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,16384,0.9068444569905599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,3072,0.010230222509966956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,16384,5.781767951117621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,16384,0.6450399822658962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,2048,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,2560,0.03084711233774821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,768,0.019923556182119582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,3584,0.022082666556040447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,7168,0.01945333348380195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,2048,0.075290666686164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,1536,0.050126221444871694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,10240,0.8070346514383951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,5120,0.010518222219414182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,3072,0.14312711026933458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,5120,0.1879324383205838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,1024,0.019143111175960965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,4096,0.10833155446582371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,6144,0.19453510973188612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,64,0.0029128889242808023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,2560,0.09011110994550918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,768,0.005247999810510212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,128,0.004488000025351842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,12288,0.04788088798522949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,51200,0.2642177740732829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,512,0.004067555483844546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,65536,0.17773600419362387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,65536,0.06976711087756686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,51200,2.3312354617648654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,3072,0.08980088763766819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,2560,0.015142222245534262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,51200,0.13414577643076578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,5120,0.00959644466638565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,12288,0.20765777428944907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,5120,0.009328000247478485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,5120,0.015896000795894198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,3072,0.03464444478352865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,51200,1.390334235297309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,6144,0.015899555550681222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,4096,1.7763919830322266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,5120,0.02423644396993849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,4096,0.01072266697883606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,2048,0.011593778100278644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,51200,1.5150586234198675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,6144,0.192138671875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,512,0.008531555533409119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,65536,0.07311555412080553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,16384,0.12061866124471028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,12288,0.9485520256890191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,51200,0.18936533398098418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,7168,5.396124521891276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,3584,0.2022764417860243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,3584,0.05292266607284546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,3072,0.1095013353559706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,10240,0.03615377677811517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,3072,0.007548444800906711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,8192,0.03814222084151374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,51200,1.7108115090264215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,51200,0.06038933330112033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,768,0.006015111174848344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,4096,0.008823999928103553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,128,0.0032435556252797446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,128,0.04241866535610623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,1536,0.006199999815887875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,6144,0.010079999764760336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,256,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,512,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,768,0.5231048795912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,65536,0.10366221931245591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,10240,0.4113502237531874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,768,0.040116445885764226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,1536,0.039388444688585066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,6144,0.017370666066805523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,3072,0.11923466788397895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,2560,0.03657244311438666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,10240,768,0.19034577740563285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,512,0.005445333404673471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,51200,1.260167121887207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,51200,51200,6.319963667127822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,6144,0.43047022819519043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,7168,0.009273777405420939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,1536,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,6144,0.023361777265866596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,4096,0.012543111211723752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,51200,0.05776800049675835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,12288,0.01688799924320645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,2560,0.028696888022952612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,5120,0.05886755386988322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,6144,0.009233777721722921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,1536,0.011187555889288584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,10240,0.20350844330257842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,6144,0.20664800537957084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,4096,0.10092800193362766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,12288,2.7897928025987415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,3584,0.010677333507272931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,1024,0.007651555869314406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,1024,0.006966222491529252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,768,0.20859289169311523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,5120,0.017690666847758822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,128,0.0064497776329517365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,7168,0.07185510794321696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,65536,0.21997422642178008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,512,0.02422488894727495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,4096,1536,0.05877600113550822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,4096,0.008393777741326226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,5120,0.33180178536309135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,8192,0.01921244462331136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,3072,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,2560,0.007547555698288812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,3584,0.07904089159435697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,10240,0.035043554173575506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,1536,0.00720177756415473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,3072,0.08432177702585857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,4096,0.061562663979000516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,32,0.0058782220714622075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,2048,0.01132266637351778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,8192,0.8708995713127984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,12288,0.18136710590786406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,7168,0.2018995549943712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,3072,0.09937066502041286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,128,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,256,0.011867555479208628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,5120,0.04540088772773743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,32,0.005316444569163852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,1024,0.004042666819360522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,512,0.01163733336660597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,256,0.012968888713253869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,8192,0.3529413276248508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,12288,0.01551999979548984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,3072,0.01462311049302419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,1536,0.014563555518786112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,1024,0.010199111368921068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,5120,0.010330666270520952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,1536,0.00943288869327969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,64,0.018025777406162687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,2560,0.025645333859655593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,65536,0.012625777886973487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,6144,0.09505689144134521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,1024,0.019659555620617337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,1536,0.05439022183418274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,4096,0.1241004467010498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,4096,0.06400622261895074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,12288,0.9985511567857531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,2560,0.13372977574666342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,2560,0.01828533411026001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,2560,0.09285155269834731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,4096,0.019500444332758587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,7168,0.031102223528756037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,7168,0.3623262246449788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,3584,0.007755555212497711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,6144,0.02311644454797109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,3072,0.006892444358931647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,128,0.006121777825885349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,16384,0.835628456539578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,3584,0.02362755603260464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,256,0.06784711281458537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,4096,0.00756888919406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,2560,0.04281066523657905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,7168,0.07166222068998548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,51200,1.340287102593316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,5120,0.02083911167250739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,1024,0.2474302185906304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,51200,0.05589333507749769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,2048,0.10550222131941055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,1024,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,2048,0.005412444472312927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,2048,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,1536,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,16384,0.02328088879585266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,1536,0.1324088838365343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,1536,0.3594817850324843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,256,0.00556355549229516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,3072,0.018844443890783522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,128,0.005382222019963794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,6144,0.0459591117170122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,64,0.0037662221325768363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,10240,0.26333512200249565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,2560,0.006108444597986009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,1536,0.026231999198595684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,8192,0.046514666742748685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,64,0.003141333245568805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,65536,0.7873848809136285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,6144,0.15793332788679335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,128,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,256,0.00869333330127928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,3072,0.010429333481523724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,5120,0.030878222650951807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,768,0.031904889477623835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,6144,0.027297778262032404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,2560,0.010135110881593492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,1536,0.007190222541491191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,128,0.13091999954647487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,3072,0.009303111169073317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,2048,0.011203555597199334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,2560,0.047888888253106006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,51200,0.8046008745829264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,6144,0.010078222387366826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,256,0.009497777455382878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,64,0.03389599919319153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,51200,0.12649067242940268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,1536,0.03249955508444045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,12288,0.058074666394127734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,10240,0.038662221696641706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,7168,0.014157333307796054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,51200,2.247889836629232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,5120,1.7111849255032008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,128,0.004616888860861461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,1536,0.011649777491887411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,2560,0.017704889178276062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,64,0.0038693332009845306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,1536,0.021402666966120403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,12288,0.22757156689961752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,3072,0.08354577753278945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,5120,0.12570844756232366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,2048,0.04632000128428141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,3072,0.059008002281188965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,6144,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,5120,0.021399110555648804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,65536,1.7576764424641926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,64,0.01824355622132619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,128,0.006140444427728653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,768,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,6144,0.02272355556488037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,6144,0.4840844472249349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,2048,0.12860355112287733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,10240,0.12476444244384766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,32,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,256,0.008411555654472774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,256,0.004769777672158347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,64,0.00425155543618732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,5120,0.9194213019476997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,4096,0.01682400041156345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,51200,0.06559733549753825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,51200,0.7480942408243815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,5120,0.031108445591396753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,8192,0.039670222335391574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,64,0.00868977771864997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,2560,0.012624888784355588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,3072,0.006060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,7168,0.12325689527723525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,10240,0.024454222785101995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,3584,0.009302222066455418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,8192,0.05112533436881172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,128,0.015615999698638916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,32,0.0026782221264309357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,256,0.0028231110837724474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,7168,0.18011289172702363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,16384,0.3854000038570828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,51200,0.13353689511617026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,1024,0.007109333243634965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,10240,0.016859556237856548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,128,0.008034666379292807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,3584,0.025119110941886902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,128,0.0027537778433826235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,2560,0.007860444486141205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,64,0.0026275554878844153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,128,0.0027253333893087176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,512,0.005416888743638992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,12288,0.1570568879445394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,6144,0.008172444171375698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,64,0.005053333524200651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,65536,0.00959288908375634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,512,0.0034702221552530923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,51200,0.4299004342820909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,16384,0.02184266679816776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,768,0.0075626663035816616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,12288,0.12297688590155707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,1024,0.009008000294367472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,6144,0.006508444332414203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,1024,0.0162764439980189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,7168,0.013387555877367655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,12288,0.04920088913705614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,768,512,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,7168,0.013368888861603208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,768,0.016694222887357075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,1024,0.0053075556125905775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,5120,0.01440177857875824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,2560,0.012326222327020435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,2048,0.006852444261312485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,10240,0.017394666870435078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,256,0.0052328887912962176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,10240,0.011137777732478248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,512,0.05643288956748115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,512,0.005511111269394557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,5120,0.03179288903872172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,2048,0.009909333454238044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,2560,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,16384,0.039686222871144615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,6144,0.023079110516442194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,64,0.0027884443600972495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,10240,4.195635477701823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,65536,0.011947555674446953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,6144,0.2326035499572754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,4096,0.006741333338949416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,4096,0.005716444303592046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,768,0.005809777726729711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,8192,0.022813333405388728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,6144,0.03899733225504557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,4096,0.009504888620641496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,51200,0.057279109954833984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,512,0.009448889229032729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,2560,0.044772442844178945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,51200,0.7933804194132487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,3584,0.3795066674550374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,7168,0.2961004575093587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,3072,0.009919111099508073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,16384,0.28547912173801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,16384,0.19195289081997344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,128,0.004594666676388847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,768,0.007654222349325816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,256,0.12947377893659803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,7168,0.03494844502872891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,768,0.01001422190003925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,4096,0.06135644515355428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,2560,0.04730044470893013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,2560,0.030062221818500098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,2048,0.03720533185535007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,6144,0.018730666902330186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,2048,0.01459111107720269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,2048,0.03772177630000644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,32,0.0027146666414207886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,65536,3.3856728871663413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,4096,0.0838728878233168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,8192,0.015987555185953777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,3584,0.016885333591037326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,768,0.012947555217478009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,3584,0.8173697789510092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,1536,0.05547822184032864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,3584,0.010106667048401302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,1024,0.00923555592695872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,5120,0.03232889042960273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,10240,0.04091644287109375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,1536,0.014314666390419006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,768,0.018023111753993563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,7168,0.6802924474080404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,2560,0.05529333485497368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,256,0.0033093332830402586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,3584,0.013991999957296582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,51200,0.19168444474538168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,64,0.005430222385459476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,1024,0.3745573361714681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,3584,0.008094222181373173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,7168,0.061852441893683545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,12288,0.01849422189924452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,65536,0.467849784427219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,6144,0.1438266701168484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,3584,0.011197333534558615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,768,0.04374577932887607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,3584,0.020299555526839364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,12288,0.030468444029490154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,32,2560,0.005535999933878581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,2560,0.011918221910794577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,32,0.007167110840479533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,2560,0.011326221956147088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,1536,0.10254933436711629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,256,0.10102755493587917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,10240,0.0348417792055342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,12288,0.021560889151361253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,32,0.006970666348934174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,10240,0.030788444810443457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,3584,0.010320000350475311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,12288,0.7923386891682943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,1024,0.0074275558193524676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,6144,0.033013333876927696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,5120,0.008127999802430471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,4096,0.017551110850440133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,6144,0.030566222137875024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,3072,0.02873599860403273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,16384,0.030468444029490154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,5120,0.015487111277050443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,2048,0.0058240000572469496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,12288,0.9322879579332141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,768,0.043162667089038424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,7168,0.024318221542570326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,16384,0.013914666242069669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1536,2560,0.00834222220712238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,4096,0.014269333746698169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,128,0.0047386665311124595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,65536,0.02647288971477085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,4096,0.025432000557581585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,2560,0.047839999198913574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,12288,0.06787644492255317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,7168,0.4729324446784125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,32,0.06085599793328179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,16384,0.31170667542351616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,12288,0.03348711133003235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,1024,0.0325564444065094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,128,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,16384,0.10466577609380086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,16384,0.0749173363049825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,16384,0.023594665858480666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,3072,0.28967732853359646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,64,0.0025439999169773525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,3072,1.4319243960910375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,1024,0.10151288906733195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,1024,0.012448888685968188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,768,0.02588266630967458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,16384,0.062407109472486705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,128,0.00387999994887246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,2048,0.04419111212094625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,2560,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,256,0.004609777695602841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,6144,0.312791109085083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,4096,0.07352711094750299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,128,0.0028106667515304354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,1024,0.008445333275530074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,4096,0.025409777959187824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,51200,0.08847822083367242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,512,0.03141422073046366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,3072,0.19407467047373453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,3584,0.03439022103945414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,8192,0.04727466570006477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,2560,0.12257156107160781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,3072,0.10816888676749335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,8192,0.2703208923339844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,1024,0.031128889984554712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,7168,0.07846933603286743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,51200,0.0735973318417867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,5120,0.023005333211686876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,7168,0.03568355573548211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,51200,0.09385422203275894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,64,0.005658666706747479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,2048,0.03740444448259141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,256,0.0032506665835777917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,64,0.007744889292452071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,8192,0.02118844456142849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,6144,0.03420711225933499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,1024,0.007697777615653143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,4096,0.6775911119249133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,128,0.0027635554886526535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,768,0.005617777920431561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,51200,1.1476639641655815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,3072,0.06051644351747301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,1024,0.0624906685617235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,7168,0.029141333368089464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,65536,2.866584989759657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,7168,0.21228355831570098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,3584,0.06897422340181139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,256,0.00509333320789867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,768,0.027967111931906805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,3072,0.10468800200356378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,1536,0.020475554797384474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,4096,0.008912000391218398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,4096,0.17753244770897758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,512,0.052185777160856456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,3584,0.010089777410030365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,256,0.0053137776752312975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,4096,1.28474669986301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,6144,0.042020443412992686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,3584,0.023569777607917786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,16384,0.04254577888382805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,2048,0.024519110719362896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,5120,0.012647999657524956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,51200,0.16490399837493896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,64,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,3584,0.02566755480236477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,256,0.021363554729355708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,2048,0.01483022173245748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,6144,0.056537780496809215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,32,0.045915553967158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,7168,0.028211557202868994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,65536,0.07167822122573853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,16384,0.047761778036753334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,10240,0.1840888924068875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,768,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,65536,0.29569866922166615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,16384,0.5066933102077907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,12288,0.14815378189086914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,32,0.006063111126422882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,8192,0.03879288832346598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,12288,0.02825066778394911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,4096,0.16731733745998809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,3072,0.049011554982927114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,128,0.006609777609507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,3584,0.012331555287043253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,10240,0.029232889413833618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,32,0.007089777953094906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,7168,0.3816995620727539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,8192,1.7390640046861436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,32,0.003137777869900068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,5120,0.08122489187452528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,8192,0.015464888678656684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,12288,1.8781102498372395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,65536,0.08411555820041233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,3072,0.010486221975750394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,2560,0.013520888984203339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,256,0.030623111459943984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,256,0.0037288889288902283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,1024,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,2560,0.035312000248167254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,1536,0.006787555499209299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,128,0.005244444641802046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,65536,0.4032275676727295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,16384,0.04560888806978861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,64,0.008593777815500895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,512,0.07184711429807875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,768,0.012123555772834353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,256,0.0327413347032335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,1536,0.010258666343159145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,32,0.029623111089070637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,7168,0.023784000012609694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,2048,0.012925333446926542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,5120,0.010812444819344414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,5120,0.02407377792729272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,2048,0.06016000111897787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,3584,0.07344800233840942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,12288,0.110152886973487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,8192,0.1490595605638292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,1536,0.004666666603750653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,8192,0.042704890171686806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,128,0.0041662222809261745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,128,0.006421333385838403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,1024,0.0148026661740409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,768,0.007622222105662028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,6144,0.09269599782096015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,8192,0.014504000544548035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,3072,0.010197333163685268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,128,0.06975199778874715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,128,0.0028328889360030494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,5120,0.19248710738288033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,12288,1.7896516587999132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,51200,0.10777688688702053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,256,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,8192,0.01701866587003072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,2048,0.028748446040683325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,6144,0.027513777216275532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,7168,0.019934222102165222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,1536,0.006816000160243776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,2560,0.11600711610582139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,7168,0.12531999746958414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,12288,0.032264888286590576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,256,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,5120,0.03164444367090861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,64,0.0447688897450765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,3072,0.01623911162217458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,3072,0.02645422187116411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,10240,0.05547733439339531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,12288,0.012639111114872826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,6144,0.014329777823554145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,4096,0.12826132774353027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,2560,0.042028443680869214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,128,0.005501333210203383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,51200,0.01292622254954444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,10240,0.0677715539932251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,256,0.2167422241634793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,2560,0.008070222205585903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,256,0.011352889239788055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,256,0.00738666703303655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,51200,8192,1.8611511654324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,4096,0.02959288822280036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,3584,0.044954667488733925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,1024,0.005921777751710679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,2560,0.6843031247456869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,768,128,0.004926222066084544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,65536,0.11472888787587483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,51200,0.06247378057903714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,12288,0.14788444836934408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,256,0.008407110969225565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,8192,0.025797333982255723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,65536,0.27622845437791615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,32,0.010122666756312052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,1024,0.006832888970772426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,51200,10.008655971950954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,65536,0.44958135816786027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,12288,0.027715555495686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,3072,0.007576000359323289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,1536,0.0053173332578606075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,2560,0.008777778181764815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,6144,0.049104889233907066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,4096,0.011329777538776398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,12288,0.016028443972269695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,7168,0.049991110960642494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,512,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,3072,0.007198221981525421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,16384,2.1166417863633895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,32,0.02681866619322035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,768,0.0059164443777667145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,51200,0.11741956075032552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,6144,0.12681510713365343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,7168,0.38991112179226345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,8192,0.2371928956773546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,256,0.0035982223020659555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,1024,0.06946755780114068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,6144,0.020993777447276648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,5120,0.021819555097156104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,8192,0.04068444503678216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,512,0.016892444756295946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,7168,0.014269333746698169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,512,0.03752266698413425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,64,0.0058515556156635284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,1536,0.016600888636377122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,10240,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,8192,0.018816888332366943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,2560,0.0753600001335144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,32,0.002667555585503578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,6144,0.023680888944202002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,10240,0.03394577900568644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,4096,3072,0.011306666665607028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,8192,0.011735999749766456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,6144,0.693145751953125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,2048,0.01352266636159685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,768,0.010778666370444827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,51200,0.30949332978990346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,64,0.0051857779423395796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,512,0.004632888982693355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,2560,0.018992000155978732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,51200,2.5196471744113498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,3584,0.43406134181552464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,8192,2.3090657128228083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,65536,0.9065280490451388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,8192,0.015399111641777886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,256,0.008089777496125963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,6144,0.05098755492104424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,5120,0.09372799926333958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,8192,0.0069297779765393995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,5120,0.011115555961926779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,64,0.005211555709441503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,1024,0.02985244327121311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,5120,0.009338666995366415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,51200,0.9880310694376627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,768,0.006385777973466449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,512,0.014488000008794995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,768,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,32,0.017275555266274344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,6144,0.03945777813593546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,256,0.028139554791980322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,3072,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,1536,0.008108444511890411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,7168,0.13870044549306235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,6144,0.013353777428468069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,256,0.0028195555011431375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,12288,1.4916968875461156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3072,512,0.0058266665372583605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,8192,0.026891556051042344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,4096,1.065800878736708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,1024,0.008104888929261101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,7168,0.01853155593077342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,128,0.0034435554924938413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,32,768,0.005724444571468566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,12288,0.0699839989344279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,1024,0.004879999905824661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,1536,0.015081778168678284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,2048,0.011630222201347351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,1024,0.005598222215970357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,65536,0.17497066656748453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,128,0.004297777596447202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2048,1536,0.01993333299954732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,65536,0.904537730746799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,16384,0.021853332718213398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,2560,0.4794933531019423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,256,0.03528799944453769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,1536,0.05137333273887634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,1024,0.05211911267704434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,2560,0.03940977652867635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,4096,0.13793067137400308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,768,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,512,0.0031022222505675424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,8192,0.018428444862365723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,1536,0.009920000202125972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,256,0.008934222161769867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,8192,0.1260746717453003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,256,0.019144000278578866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,768,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,65536,0.0808124476008945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,65536,0.09555288818147446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,12288,0.122871994972229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,51200,2.284589343600803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,64,0.02905333373281691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,1536,0.036271111832724676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,128,0.004886222382386525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,5120,1.8373546600341797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,7168,0.10551289055082534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,5120,0.18993066416846383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,3584,0.02574488851759169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,3584,0.030412445465723675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,51200,0.19803555806477866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,3072,0.022032888399230108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,8192,0.10641244384977554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,1536,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,512,0.004104888687531154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,10240,0.08601511187023586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,10240,0.01905866629547543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,1024,0.02276533345381419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,1024,0.03308444552951389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,3584,0.008601778083377415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,1024,0.006609777609507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,5120,0.04182577795452542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,10240,0.14633599917093912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,3584,0.16576977570851645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,3072,0.007133333219422235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,2048,0.07309422228071424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,256,0.0469084448284573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,8192,0.017856000198258292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2048,16384,0.332202672958374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,10240,0.07602666483985053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,12288,0.12435732947455512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,7168,1.0463253657023113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,7168,0.028877334462271795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,8192,0.024541333317756653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,64,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,51200,0.3318675623999702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,7168,0.06278310881720649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,10240,1.8850169711642797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,512,0.011568889021873474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,5120,0.017815111411942374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,10240,0.1653555499182807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,8192,0.03190844588809543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,32,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,32,0.1627982192569309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,3584,0.08934577968385483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,51200,1.3824657864040797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,1024,0.01761511133776771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,1536,0.004726222405831019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,3072,0.007532444265153673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,4096,0.14289955298105875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,4096,0.016322667400042217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,256,0.006596444381607904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,5120,0.08307022518581815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,512,0.005473777651786804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,4096,0.008778666456540426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,3072,0.05618400043911404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,51200,0.6896062427096896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,3584,0.016717332932684157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,10240,0.03841244512134128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,2560,0.008163555628723567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,7168,0.03885511226124234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,8192,0.6955875290764703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,51200,768,0.48775646421644425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,64,0.007733333441946242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,3072,0.8662186728583442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,1536,0.011352000137170156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,1536,0.016576000385814242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,65536,2.9695012834337025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,6144,0.24993244806925455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,512,0.010558221903112201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,3072,0.02884977724817064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,3584,0.038648890124426946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,3584,0.02103288968404134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,51200,0.14497333102756077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,3072,0.04990577697753906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,3584,0.028637333048714533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,1024,0.0058053334554036455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,8192,0.007096889118353526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,2560,0.13349955611758763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,65536,1.8271876441107855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,768,0.015306666493415833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,5120,0.13442933559417725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,3584,0.21771021684010824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,51200,1.3195546468098958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,3072,0.03202222122086419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,2560,0.46902041965060765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,6144,0.06891466511620416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,1536,0.021198223034540813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,7168,0.012640000217490725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,128,0.025796444879637823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,4096,0.013749333719412485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,12288,0.09752444426218669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,768,0.004243555582231945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,5120,0.03396000133620368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,10240,0.04050666756100125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,10240,0.21242845058441162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,6144,0.07729688617918226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,5120,0.13498844040764704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,1536,0.015951111912727356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,4096,0.011550222006109027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,16384,2.415156470404731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,8192,0.02809422214825948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,10240,0.00942133367061615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,10240,0.19711111651526558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,2560,0.020569778150982328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,8192,0.016113777955373127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,10240,0.04795555604828728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,7168,0.011205332974592844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,8192,0.021202666891945735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,7168,1.0869048436482747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,4096,0.1371199952231513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,2560,0.03969600134425693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,2560,0.006692444284756978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,8192,0.24078310860527885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,5120,0.012311110893885294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,1024,0.09629155529869927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,3584,0.05898844533496433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,1536,0.004416888786686791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,1536,0.011213333242469363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,768,0.020961778031455144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,16384,0.030508445368872747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,65536,0.2651093271043566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,3072,0.017071111334694754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,51200,0.3095022307501899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,768,0.005662222289376789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,128,0.04214488797717624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3072,1536,0.006423111177153057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,5120,0.006607111129495833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,7168,0.014892444014549255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,32,0.04303555687268575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,2560,0.007181333170996771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1536,51200,0.09012533558739556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,8192,0.02688177757793003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,2560,0.00775111135509279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,4096,3.371797349717882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,10240,0.061384889814588756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,2048,0.007092444433106317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,2560,0.009088888764381409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,51200,0.10994222429063584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,768,0.05232622226079305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,51200,149.444091796875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,2048,0.06424178017510308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,768,0.30038756794399685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,64,0.0028408887899584244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,768,0.01186044431395001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,5120,0.06141689088609484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,16384,0.10150400135252212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,768,0.004060444318585926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,1024,0.06941422488954332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,4096,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,16384,0.10573422246509129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,7168,0.11466844876607259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,1536,0.030864887767367895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,2048,0.00867377801073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,16384,0.9257617526584201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,5120,0.7191502253214518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,6144,0.03177244464556376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,16384,0.09358400106430054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,2048,0.015239111251301236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,768,0.025574222207069397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,7168,0.02411288850837284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,768,0.004014222158326043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,256,0.023355555203225877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,2048,0.0166302224000295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,65536,0.27883466084798175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,2560,0.009574222067991892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,7168,0.015239111251301236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,512,0.0036524443162812125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,2560,0.007853333320882585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,16384,0.053040888574388295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,65536,2.551028357611762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,6144,0.5988017717997233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,512,0.019710222880045574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,10240,0.19075822830200195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,1024,0.005993777679072485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,32,0.025263110796610515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,51200,0.1929288837644789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,4096,0.012641777594884237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,6144,0.017440888616773818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,2560,0.043904887305365674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,2560,0.008839111361238692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,3072,0.01683999929163191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,16384,3.5303368038601346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,512,0.00941422250535753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,3072,0.016699555847379897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,4096,0.06053600046369764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,6144,0.44026488727993435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,8192,0.01735466718673706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,1024,0.00812177773978975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,1536,0.008707555631796518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,256,0.0052728888889153796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,2560,0.030934221214718286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,512,0.00942311104800966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,12288,0.2763662338256836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,3584,0.008954666554927826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,12288,0.05875911315282186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,256,0.0032720000793536506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,3584,0.008649778034951951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,2560,0.06948622067769368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,3072,0.015841777125994366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,12288,0.4326017697652181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,10240,0.07304888963699341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,12288,0.36607911851671004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,1536,0.0136835558546914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,65536,6.617394765218099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,1536,0.007846222155623967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,512,0.004757333132955763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,256,0.013824889229403602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,6144,0.3950933350457086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,1536,0.04691288868586222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,2560,0.013602666556835175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,65536,0.07428622245788574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,3584,0.01886666648917728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,1024,0.04102222124735514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,10240,0.20561688476138643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,4096,0.17801154984368217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,256,0.005514666852023866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,128,0.003096888876623578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,256,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,2560,0.020888888173633154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,10240,0.026185777452256944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,1536,0.006246222390068903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,8192,0.023189332750108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,2048,0.007144889069928064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,65536,0.1281013356314765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,10240,0.18544088469611275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,7168,8.884808858235678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,6144,0.08212444517347547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,6144,0.062330669826931424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,256,0.009387555221716562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,3072,0.027865777413050335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,10240,0.2144977781507704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,2560,0.012585777375433179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,3584,0.01826755536927117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,128,0.00600888869828648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,8192,0.27979644139607746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,1024,0.012823999755912356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,12288,2.7197413974338107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,2560,0.07607733541064791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,3584,0.07664622200859918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,128,0.00720177756415473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,3072,0.043064888980653554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,51200,0.05598044395446777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,1024,0.004281777888536453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,16384,0.3545200030008952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,2560,0.0861368907822503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,65536,0.7425120141771105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,2560,0.008550222549173567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,64,0.006722666737106111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,1536,0.009597332941161262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,4096,0.010168888502650792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,7168,0.01641155613793267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,1536,0.0053591111467944244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,51200,0.7428799735175239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,256,0.014712888333532544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,8192,0.014727110664049784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,3584,0.00813244448767768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,1536,0.007496000164084964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,3584,0.011375111010339526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,12288,0.023131555981106226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,16384,0.04882755544450548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,2048,0.07604177792867024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,3072,0.00867022242810991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,1536,0.03599377804332309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,8192,2.983571582370334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,2560,0.03674222363366021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,16384,0.05620266993840536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,10240,0.03035022152794732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,3072,0.02587466604179806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,8192,0.022820444570647344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,128,0.0028968888024489083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,4096,0.15965421994527182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,6144,0.12075200345781113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,2560,0.022836445106400385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,5120,0.028613332245084975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,5120,0.011712888876597086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,2560,0.011807111402352652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,64,0.002510222295920054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,4096,0.01072533345884747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,3584,0.017258667283587985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,5120,0.2630959881676568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,6144,0.04668000009324816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,128,0.3198106553819445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,1536,0.006160888820886612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,768,0.0058515556156635284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,1536,0.006539555473460092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,4096,0.07368177837795682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,768,0.01795111099878947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,51200,0.29384533564249676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,32,0.00658044425977601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,7168,0.11210044225056966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,2560,0.01570933394961887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,65536,1.0191884570651584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,64,0.003362666608558761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,10240,0.024919110867712233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,16384,0.027186666925748188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,65536,0.13595467143588597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,256,0.007337777978844113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,3072,0.012215110990736218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,8192,0.12034489048851861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,2048,0.005473777651786804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,5120,0.31766756375630695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,1536,0.006688000013430913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,10240,0.043041778935326465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,51200,0.581032911936442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,128,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1024,32,0.00629688882165485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,3584,0.012907555533779992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,3072,0.007848000360859765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,7168,0.2633599970075819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,3584,0.010273777776294285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,8192,0.050364444653193154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,256,0.00536088893810908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,256,0.004582222137186262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,6144,0.6552497545878092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,2560,0.6709280014038086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,16384,0.5673564275105795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,4096,0.0092257774538464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,1024,0.004324444466167026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,32,0.008031999899281396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,4096,1.9797982109917536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,12288,0.04196622305446201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,7168,0.12514132923550075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,2048,0.018205333087179396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,32,0.005592000153329637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,65536,0.1473884450064765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,65536,2.09879027472602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,2048,0.016480889585283067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,51200,7.44777594672309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,1024,0.041314668125576444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,3072,0.03949777616394891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,16384,0.025390222668647766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,128,0.021758221917682227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,51200,0.06568533182144165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,51200,0.5382311079237196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,65536,0.09581244654125637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,3584,0.02681866619322035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,65536,19.721600002712673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,256,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,768,0.04330133398373922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,12288,0.03735111157099406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,256,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,1024,0.025514667232831318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,512,0.008692444198661381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,1536,0.012307555311255984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,51200,0.5997449027167426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,12288,0.02315555512905121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,8192,0.11202755239274766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,2048,0.019646222392718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,3584,0.009094222552246517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,1536,0.008068444828192392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,512,0.006055999961164262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,128,0.009490666290124258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,256,0.015950222810109455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,2048,0.04545955525504219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,512,0.06462577978769939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,10240,0.42159467273288304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,1536,0.32091111607021755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,128,0.010508444574144153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,51200,0.06524178054597643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,1536,0.006626666833957036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,3072,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,6144,0.02438933319515652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,512,0.008815999660227034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,6144,0.14207200209299722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,1024,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,51200,0.14248889022403294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,2048,0.36467732323540586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,5120,0.02405333353413476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,64,0.004722666823201709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,1536,0.017398221625222098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,3584,0.1548337803946601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,10240,2.98597780863444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3072,256,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,512,0.014344000154071383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,5120,0.03418755531311035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,128,0.003015999992688497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,7168,0.006773333168692059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,32,0.003354666754603386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,3584,0.008422222402360704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,2560,0.09223644600974189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,65536,8.246683756510416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,4096,0.03891288903024461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,65536,0.08113511403401692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,6144,0.008119111259778341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,4096,0.04456088940302531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,1536,0.014780445231331719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,5120,0.3815457820892334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,10240,0.1712062226401435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,3072,0.3260364532470703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,32,0.014182221558358936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,256,0.005295111073387994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,256,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,2560,0.007810666329330868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,64,0.003040888864133093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,256,0.005771555420425203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,512,0.008619555168681675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,3072,0.1425457795461019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,5120,0.022447999980714586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,1536,0.004882666799757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,12288,0.2209120061662462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,768,0.008776889079146916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,2048,0.7266409132215711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,1024,0.018408889571825664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,65536,0.7613564597235786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,16384,0.3169911172654894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,5120,0.029883556895785864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,3584,0.010490666660997601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,51200,0.05608533488379585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,1024,0.22079911496904162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,768,0.016495111915800307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,32,0.002410666603181097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,256,0.04590577880541483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,8192,0.15623911221822104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,128,0.006236444330877728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,3072,0.17011377546522352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,5120,0.03314044409328037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,7168,0.13018843862745497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,2560,0.010467555787828235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,16384,0.06008355485068428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,4096,0.009000889129108852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,8192,0.015290665957662793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,1024,0.17876622411939833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,4096,0.012732444538010491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,12288,2048,0.027607111467255488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,6144,0.04038044479158189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,32,0.005981333139869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,5120,1.6366933186848958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,6144,0.08721511231528388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,16384,0.9221430884467231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,8192,0.04135466615358988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,512,0.013443555268976422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,1536,0.006351999938488007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,768,0.016381333271662395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,6144,0.38021334012349445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,8192,0.03083555565940009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,2048,0.014255111416180929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,2560,0.016355555918481614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,7168,0.019267555740144517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3584,51200,0.0959671139717102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,768,0.006559111177921295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,10240,0.050809777445263334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,3072,0.027821332216262817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,1024,0.02442666722668542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,2048,0.0069724445541699724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,10240,0.006544000158707301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,8192,0.012279999752839407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,4096,0.07229955328835382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,6144,0.02416177756256527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,7168,0.16923911041683623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,256,0.04605066776275635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,8192,0.0207022229830424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,3584,0.007528000407748752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,128,0.004404444247484207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,1536,0.020023110840055678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,768,0.02238044473859999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,1024,0.006138666636413998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,12288,0.017695110705163743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,8192,0.021839110387696162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,3584,0.023464888334274292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,65536,2.2989156511094833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,256,0.006498666687144174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,10240,0.01837688850031959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,2560,0.07254488600624932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,128,0.007664889097213745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,5120,0.17361688613891602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,1536,0.1931448909971449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,12288,0.02037333283159468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,51200,0.013453333742088742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,12288,0.020310221446885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,2048,0.04059022333886888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,10240,0.06670133272806804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,1024,0.13446044921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,1024,0.07361688878801134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,1536,0.010967111421955956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,65536,1.7679902182685003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,65536,0.27813956472608775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,256,0.017295110556814406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,128,0.012395555774370829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,1536,0.07161688804626465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,3072,0.022661333282788593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,65536,512,0.37491467263963485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,2560,0.013615999784734515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,12288,0.031001776456832886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,3072,0.8816844622294108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,8192,0.08502755562464397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,128,0.005283555636803309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,6144,1.3884666230943468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,2048,0.011941333611806234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,10240,0.03894933395915561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,10240,0.13915110958947077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,128,0.006191111273235745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,3584,0.009600888523790572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,8192,0.21549777189890543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,16384,0.012261333564917246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,2560,0.006124444305896759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,512,0.005524444497293896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,4096,0.014095111025704278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,10240,512,0.015244444211324057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,16384,0.02239022155602773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,3584,0.0471111105548011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,10240,0.6797599792480469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,12288,0.00904088881280687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,7168,0.043565332889556885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,10240,0.024940444363488093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,7168,0.014473777678277759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,8192,0.016032889485359192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,1536,0.03924355573124356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,2560,4.6524810791015625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,51200,4.071080101860894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,8192,1.0830728742811415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,768,0.008822222550710043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,51200,2048,0.03982222080230713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,2560,0.007837333612971837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,5120,0.05256888932651944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,10240,0.024314666787783306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,7168,0.012179555164443122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,3584,0.027487110760476854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,512,0.010032888915803697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,128,0.0188711103465822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,1024,0.006647111227114995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,6144,0.04734666811095344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,128,0.01146577795346578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,10240,0.1875306632783678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,128,0.027986665566762287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,65536,0.23985245492723253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,65536,0.10285600026448567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,8192,0.024138665861553613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,3584,0.10278044144312541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,512,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,2560,0.14457689391242132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,3072,0.049307554960250854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,1536,0.08366933133867051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,16384,29.324839274088543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,128,0.010467555787828235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,7168,0.029299557209014893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,12288,0.022282666630215112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,1024,0.012792000340090858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,256,0.0034719999465677473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,51200,1.3708506690131292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,65536,0.25075022379557294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,12288,2.036810768975152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,7168,0.7078142166137695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,51200,1.7105724546644423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,65536,0.08113688892788358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,768,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,7168,0.06995733578999837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,4096,0.9493822521633573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,7168,3.2696694268120656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,128,0.005054222212897407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,1024,0.00461777796347936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,2048,0.06629688872231378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,512,0.007843555675612556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,51200,6.2947510613335504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,12288,0.7382488780551486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,3584,0.025011556016074285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,65536,0.42586223284403485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,7168,0.06159289015663994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,5120,0.01274222218328052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,128,1024,0.029855999681684706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,4096,0.015744000673294067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,64,0.0030497778207063675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,3584,0.009288888838556077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,2048,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,3584,0.012267555627557965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,6144,0.041448887851503156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,65536,0.21339555581410727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,2048,0.020729778541458976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,2560,0.15659111075931126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,1024,0.010242666635248397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,1536,0.17572800318400064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,128,0.0074924445814556545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,51200,5.233194563123915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,32,0.009486222432719337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,4096,0.010640888578361936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,5120,0.008868444297048781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,5120,0.1688986619313558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,1024,0.04550844430923462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2048,65536,0.08066044251124065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,32,0.004192000048028098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,512,0.003375111147761345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,128,0.012150222228633033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,6144,0.17007644971211752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,2560,0.007151111132568783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,32,0.002513777878549364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,3584,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,1024,0.013629333012633853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,1024,0.022273777259720698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,1536,0.012268444730175866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,1024,0.0053431110249625305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,256,0.013960000541475085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,10240,0.24609867731730142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,2048,0.30792445606655544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,12288,0.025386666258176167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,1536,0.03590133455064561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,51200,0.30852532386779785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,512,0.005347555710209741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,8192,0.06469066937764485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,64,0.0034959999223550162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,512,0.041082666979895696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,3072,14.856121487087675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,768,0.02590488890806834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,7168,0.23101422521803114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,256,3072,0.026199999782774184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,12288,0.40454755889044863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,512,0.008404444489214156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,65536,512,0.020961778031455144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,2560,0.007856888903511895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,3072,1.7248382568359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,128,0.04200889004601372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,3584,1.8159137301974828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,8192,0.044341332382626004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,7168,2.9258374108208547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,3072,0.7200737529330783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,256,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,6144,0.1299226681391398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,5120,0.2601315445370144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,10240,0.018632888793945312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,512,0.011669333610269757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,3072,0.15025066004859075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,1536,0.011504888534545898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,6144,0.010436444646782346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,8192,0.13219822777642146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,3584,0.02349866595533159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,128,0.003015111097031169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,128,0.004703999807437261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,51200,0.11551466253068711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,12288,0.04254755708906385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,1024,0.056877334912618004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,768,0.004036444342798657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,256,1536,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,256,0.006778666542636023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,1024,0.006702222343948152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,3072,0.037674665451049805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,8192,0.036159111393822566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,128,64,0.002512000087234709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,2560,0.043600890371534556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,32,0.002977777686383989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,6144,0.018145778112941317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,3584,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,51200,1.3445332845052083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,5120,0.08186222447289361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,5120,0.10434399710761176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,1536,0.014513777361975776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,768,0.017208000024159748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,1024,0.17114310794406465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,4096,0.00979644474056032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,16384,51200,8.824868943956163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,8192,0.3535768985748291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,2560,0.03552888830502828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,10240,0.0075093333919843035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,7168,0.04158133268356323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,16384,0.04063377777735392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,10240,0.2615617646111382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,128,0.004328000048796336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,32,0.007280000381999546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,1024,0.01239111108912362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,10240,0.03956088754865859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,768,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,2048,0.007278222176763747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,1536,0.007600000335110559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,3072,0.02455733385350969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,3584,0.12943022780948216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,3072,0.008620444271299574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,1024,0.011890666352377998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,10240,0.12756800651550293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,512,0.004421333471934001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,3584,0.013053333593739403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,2048,0.4360195530785455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,16384,0.026730666557947796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,2048,0.0911102228694492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,3584,0.12882310814327663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,8192,0.01460533340771993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,128,0.13140800264146593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,3584,0.056151111920674644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,128,0.0046453335218959385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,3072,0.012243555651770698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,16384,0.04823822114202711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,3584,0.017391110459963482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,7168,0.029459555943806965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,2048,0.029363556040657893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,65536,0.07401333252588908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,65536,0.2228355672624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,8192,1.2631200154622395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,512,0.007227555745177799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,64,0.0057608890864584185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,256,0.003590222034189436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,12288,0.21631733576456705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,32,0.08199289109971789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,4096,0.6945777999030219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,512,0.04846310946676466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,3072,0.011759999725553723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,51200,0.11274045043521458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,4096,0.09030310975180732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,512,0.026525333523750305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,128,0.01310400002532535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,3584,0.02718044486310747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,5120,0.00998044427898195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,1024,0.008168000313970778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,7168,0.19427733951144746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,12288,0.029684444268544514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,4096,0.12987911701202393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,256,0.01961955593691932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,512,0.005123555660247803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,2560,0.012340444657537671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,65536,0.10328533252080281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,1536,0.037941333320405744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1024,7168,0.014283556077215405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,10240,0.16281244489881727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,8192,0.047500444783104785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,51200,0.8586613337198893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,4096,0.005924444645643234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,65536,768,0.023557333482636347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,10240,0.05931378073162503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,10240,0.013506666653686099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,768,0.027502222193611994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,3072,0.005928888916969299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,1024,0.416357331805759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,4096,0.013569778038395775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,2560,0.014923555983437432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,1536,0.010000000397364298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,10240,3.210054185655382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,3072,0.011659555964999728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,128,0.02130488885773553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,1536,0.028243554963005915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,768,0.00425155543618732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,64,0.0074720001882976955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,128,0.0046533333758513136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,10240,0.29664622412787545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,6144,0.058373332023620605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,16384,3.085016886393229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,4096,0.014704000618722705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,65536,5.986358218722874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,65536,1.0147590637207031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,6144,0.42636622322930234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,12288,0.05521777934498257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,10240,0.017687110437287223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,128,0.006331555545330048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,51200,0.062330669826931424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,2560,0.11812444527943929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,32,0.025221332907676697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,3584,0.06553066439098783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2048,512,0.005092444519201915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,256,128,0.0026604444202449587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,1536,0.04253155655331082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,256,0.005039111193683413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,3072,0.008238222036096785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,5120,0.01347555551264021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,2560,0.006641777853171031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,51200,0.13671999507480198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,65536,0.13275910748375788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,8192,0.047136889563666455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,5120,0.3616666793823242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,64,0.007439999944633908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,3584,0.02494844463136461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,1024,0.008768888811270395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,51200,0.11947200033399795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,8192,0.21869422329796684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,256,0.012999110751681857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,4096,0.11512445078955756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,768,0.027535999814669292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,51200,1.1535626517401802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,5120,0.6373262405395508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,8192,0.05619821945826212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,3072,0.04699644446372986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,6144,0.7412586741977267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,256,0.04825333423084683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,7168,0.0343137780825297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,3584,0.018838221828142803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,128,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,7168,0.017539555827776592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,768,0.005960888746711943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,4096,0.08340177933375041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,768,0.018743111027611625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,65536,0.5796213150024414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,128,0.06904888815350003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,32,0.012709333664841123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,5120,0.0936737789048089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,64,0.0028568889117903183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,64,0.005060444275538127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,5120,0.021367111139827307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,3584,0.16180622577667236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,6144,0.4539208941989475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,3072,0.016606221596399944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,768,0.02053155501683553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,7168,0.025493333737055462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,1024,0.004879999905824661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,51200,0.2706604533725315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,65536,0.4213324387868245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,3584,0.10097244713041519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,32,0.00810933361450831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,768,0.005696889013051987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,256,0.10564088821411133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,5120,0.01458666721979777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,5120,0.009167110754383935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,64,0.022582223018010456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,8192,0.5078995492723253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,3072,0.7971804406907824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,5120,0.013973333769374423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,3072,0.03030666708946228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,6144,0.038839111725489296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,6144,2.2081021202935114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,65536,0.1019235584470961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,1024,0.026564444104830425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,64,0.015955555770132277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,2048,0.3600746790568034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,51200,9.859936184353298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,16384,0.08021866612964206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1024,51200,0.38495643933614093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,256,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,768,0.009768889182143742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,64,0.005756444401211209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,1536,0.05208622084723579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,3584,0.01125955581665039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,1536,0.012127111355463663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,5120,0.006666666517655055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,3584,0.6754284434848361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,4096,0.014456000592973499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,2048,0.5477706591288248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,64,0.007859555383523306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,4096,0.02308711078431871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,3584,0.13387378056844076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,12288,0.010595555934641095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,3072,0.06521155436833699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,65536,0.07413955529530843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,10240,3.707902272542318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,32,0.006452444526884291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,3072,0.01944177846113841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,128,0.003642666671011183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,2560,0.0851866669125027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,512,0.0069253332912921906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,2048,3.340685314602322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,3584,0.09504089090559219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,51200,5.819394853379991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,32,0.005165333135260476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,65536,0.1332719988293118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,1024,0.04005511270629035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,512,0.007986666427718269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,512,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,51200,0.05501866671774122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,64,0.00553688903649648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,4096,0.01555288831392924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,51200,0.1743253337012397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,16384,0.06222044759326511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,768,0.004138666722509596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,10240,0.006667555620272954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,51200,0.09814044502046372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,3584,0.00906577789121204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,2048,0.03414222266938951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,6144,0.05189777745140923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,5120,0.2313715616861979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,3072,0.016707556115256417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,1024,0.003925333420435588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,8192,0.01553066737122006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,6144,0.02571733295917511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,16384,0.07931555642022027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,12288,0.46057865354749894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,768,0.15086311764187282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,1024,0.0041599998043643106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,5120,0.018446221947669983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,4096,0.012793777717484368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,4096,0.013109332985348172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,7168,0.4373937712775336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,51200,0.2583199871910943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,2048,0.09849955638249715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,1024,0.004057777838574516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,128,0.00852177788813909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,128,0.005689777847793367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,65536,0.6613849004109701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,2048,0.007623111208279927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,3072,0.024414221445719402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,1536,0.01960000064637926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,65536,0.05910933017730713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,1024,0.03802400165134006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,32,0.0034008889148632684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,768,0.13337511486477321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,7168,0.01090222183201048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,2048,0.011209777659840055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,65536,0.23222043779161242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,256,0.006130666782458623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,6144,0.23261600070529512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,1536,0.02329066726896498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,51200,0.3785582118564182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,4096,0.008325333396593729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,5120,0.012268444730175866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,2048,0.01717333330048455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,12288,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,65536,0.07616355684068468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,2048,0.3020915455288357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,512,0.004554666578769684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,16384,0.007568000091446771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,512,0.0033866665843460294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,3072,0.03419822123315599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,8192,0.03908266623814901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,3072,0.6418568823072646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,768,0.05375022358364529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,256,0.011704000333944956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,64,0.007582222421964009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,128,0.005068444543414646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,7168,0.10310844580332439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,1024,0.005903111149867375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1536,12288,0.026872000760502283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,1024,0.014118222726715935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,1024,0.005026666654480828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,2560,0.06772355238596599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,1024,0.034443553951051496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,4096,0.020733333296246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,51200,0.6310764418707954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,8192,0.2843502097659641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,65536,0.06778133577770658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,5120,0.02106488909986284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,6144,0.04598844382497999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,3072,0.048559112681282886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,51200,0.20332088735368517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,7168,0.059566219647725425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,1536,0.0064293332397937775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,3584,0.09673600064383613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,10240,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,51200,0.12302222516801621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,7168,0.05605777766969469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,3072,0.04448622133996752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,256,0.00665244460105896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,768,0.005744888964626525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,128,0.0041955556306574075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,3072,0.02500177754296197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,512,0.025773333178626165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,51200,0.06209777461157905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,8192,0.5625244246588813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,1024,0.006138666636413998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,6144,0.015979554918077257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,2560,1.0980763965182834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,3072,0.009832888841629028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,2048,0.06368088722229004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,16384,0.44716623094346786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,1536,0.009321778184837764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,128,0.0027804445061418745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,128,0.004769777672158347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,32,0.00282222218811512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,7168,0.011022222538789114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,10240,0.019448000523779128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,16384,3.4005680084228516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,256,0.009669333696365356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,51200,0.33094310760498047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,1024,0.016591111818949383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,3584,0.014704888065656027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,3584,1.27928892771403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,16384,0.15587822596232095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,12288,0.5498702261182997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,2560,0.009783999787436591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,2560,0.9438275231255425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,512,0.005294222384691238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,2560,0.018453333112928603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,768,0.007296000089910295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,65536,0.19753866725497773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,16384,1.1158711115519206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,3584,0.04001155495643616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,3584,0.0177182224061754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,768,0.02091022166940901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,12288,0.02958933181232876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1536,2048,0.05451822280883789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,10240,0.025097777446111042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,3072,0.01825777855184343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,256,0.005438222239414851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,8192,0.04314666655328539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,32,0.012103111379676394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,7168,0.08978844351238674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,768,0.005397333453098933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,2048,0.0075786668393347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,65536,2.2748968336317277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,8192,0.009927999642160203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,1024,0.04154933492342631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,3072,0.006771555377377405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,64,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,3584,0.010320000350475311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,2560,0.007169778148333232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,32,0.05125955409473843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,51200,2.9526123470730252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,2560,0.11392354965209961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,7168,0.26197955343458385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,512,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,256,0.0037546666959921518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,5120,0.013942221800486246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,16384,0.3116524484422472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,8192,0.24847645229763457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,51200,1.4786906772189672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,5120,0.01905066602759891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,3072,0.28679998715718585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,4096,0.08192444509930082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,8192,0.016193777322769165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,2560,0.008174222376611497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,51200,0.016536888149049547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2048,256,0.0047493332790003884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,16384,4.036312103271484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,12288,0.6397075653076172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,5120,1.0207049051920574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,10240,0.018778666853904724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,768,0.011328000161382886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,64,0.006631111105283101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,256,0.006630222416586346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,3584,0.009030222064918941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,256,0.005623111294375525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,1536,0.2205022176106771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,5120,0.01811288959450192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,16384,0.010974222587214576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,2560,0.13361422220865884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,3584,0.1313635508219401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,2048,0.13507378101348877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,3072,0.0599306689368354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,65536,9.336695353190104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,7168,0.4263759983910455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,7168,0.08188711272345649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,4096,0.023542222049501207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,2048,0.009285333255926767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,32,0.009133333133326637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,16384,1.123204443189833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,2560,0.11087199714448716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,12288,0.09711733129289414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,2560,0.007089777953094906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,12288,0.28111555841234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,1536,0.02109688851568434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,6144,0.18022578292422822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,8192,0.01789688898457421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,4096,0.08821688758002387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,2560,0.025645333859655593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,10240,0.329090674718221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,512,0.02360444433159298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,12288,0.0648897753821479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,256,0.0062773335311147906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,51200,0.1238666640387641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,12288,0.008904889225959778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,6144,0.02404800057411194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,64,0.027836445305082533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,65536,2.0321458180745444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,2048,0.8585564295450846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,256,0.0447671115398407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,2048,0.018830221560266282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,512,0.004879999905824661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,128,0.006485333459244833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,5120,0.0922657781177097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,12288,0.0069146665434042616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,12288,0.019298666053348117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,3584,0.009010666774378883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,8192,0.023176888624827068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,64,0.010842666857772403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,1536,0.01091822236776352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2048,2048,0.05252710978190104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,65536,0.41701510217454696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,256,0.00646577775478363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,12288,0.037455999188952975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,3584,0.08172800143559773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,3072,0.21737866931491426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,128,0.011772444678677453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,51200,0.09629155529869927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,51200,0.06645422511630587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,51200,0.07404355870352851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,3584,0.03558222121662564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,4096,0.020063110523753695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,10240,0.03231733375125461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,768,0.007111999723646376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,1536,0.008007110820876228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,51200,0.13263821601867676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,6144,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,256,0.005142222262091107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,32,0.04242577817704943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,8192,0.09953422016567653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,4096,0.0377511117193434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,3584,0.006834666762087081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,51200,0.05650399790869819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,12288,0.14199288686116537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,16384,0.03927288783921136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,512,5120,0.0074471111098925276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,2048,0.014725334114498563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,256,0.00546577779783143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,10240,65536,0.5208924611409506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,7168,0.015198222464985318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,2048,0.013150222599506378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,5120,0.022314666046036616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,8192,0.11504267321692573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,12288,0.039248890346950956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,12288,0.050534221861097545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,768,0.0034773333205117118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,3072,0.018441778090265062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,51200,0.6198035346137153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,768,0.005607111172543631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,65536,0.02479377720091078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,3072,0.023754666248957317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,32,0.010453333457310995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,6144,0.011608888705571493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,64,0.028911110427644517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,8192,0.016161777906947665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,10240,0.37308888965182835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,16384,0.7122666570875379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,2048,0.22811378373040092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,128,0.004095111042261124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,3072,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,7168,0.024904888537194993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,512,1536,0.0058408888677755995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,3584,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,10240,0.09972711404164632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,128,0.004046222401989831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,12288,0.1527680026160346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,12288,0.035980443159739174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,256,0.031868444548712835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,32,0.038894222842322454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,2048,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,3072,0.02165777815712823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,4096,0.014871110518773397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,5120,0.8591359986199273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,8192,0.17358756065368652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,512,0.0170151111152437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,2048,0.00850933293501536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,3584,0.5627173317803277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,256,0.004034666551484002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,128,0.031876444816589355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,7168,3.0654089185926647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,32,0.006280888699822956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,1024,0.008255999949243333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,4096,0.010563555690977307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,6144,0.23271467950608996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,512,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,12288,0.033424887392256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,3072,0.7147955364651151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,256,0.00565244464410676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,1536,0.013844444519943662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,64,0.022912888063324824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,6144,3.4425697326660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,1024,0.01849422189924452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,51200,0.18316711319817436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,512,0.004338666796684265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,1536,0.08524533112843831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,12288,0.04820177621311612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,256,0.0311155551009708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,51200,0.2623431152767605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,12288,0.16882310973273384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,6144,0.013194666968451606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,6144,0.011935999823941125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,3584,0.009875555833180746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,1024,0.010580444501505958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,4096,0.044414222240448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,8192,0.05737511316935221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,3584,0.0243511117166943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,3584,0.01092888911565145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,1024,0.004006222304370669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,2560,0.23010132047865126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,1536,1.0074052810668945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,256,0.005623111294375525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,7168,0.4399857785966661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,3584,0.009575111170609793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,64,0.015041778484980265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,256,0.0046977777447965415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,768,0.007160888777838812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,65536,0.08635022242863973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,256,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,65536,3.8375066121419272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,12288,12288,0.7649555736117892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,64,0.010455999937322406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,512,0.176893326971266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,1536,0.06699555450015597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,7168,0.15299377176496717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,2048,0.0074933336840735535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,8192,0.12744800249735513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,65536,0.3687137762705485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,16384,0.04783466789457533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,3072,0.03130311104986403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,16384,0.07234488593207465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,7168,0.06519999768998888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,3584,0.01741066740618812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,128,0.0053048887186580235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,1024,0.007787555456161499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,2048,0.011339555184046427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,6144,0.019336889187494915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,16384,0.04545688960287306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,3584,0.008839999636014303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,1024,0.022056889202859666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,16384,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,12288,0.0656800005171034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,51200,1.3984852896796331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,8192,0.02071644365787506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,4096,0.02994399931695726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,4096,0.01887022289964888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,8192,0.023052444060643513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,768,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,4096,0.011528000235557556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,51200,1.5894533793131511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,5120,0.18522400326199004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,7168,0.023287110858493384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,1024,0.007592000067234039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,512,0.028015110227796767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,512,0.11915911568535699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,3584,0.008032889001899296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,16384,0.05801066425111559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,3072,0.03976710968547397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,512,0.0572684473461575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,256,0.005349333501524395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,51200,0.06673333379957411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,65536,0.35452977816263836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,4096,0.5882826911078559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,2560,0.08242044183943006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,7168,0.016137777103318107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,6144,0.37831465403238934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,1536,0.14830311139424643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,3584,0.008950221869680617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,768,0.0348639984925588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,5120,0.25945599873860675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,16384,0.9161315494113498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,51200,1.3102373547024198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,1536,0.006598222172922558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,768,0.006736889067623351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,6144,0.011599111060301462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,65536,40.440938313802086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,8192,0.0395760006374783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,4096,0.008239111138714684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,64,0.006548444430033366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,16384,0.14061689376831055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,6144,0.06853155295054118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,2560,0.04580533173349169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,12288,0.08295022116767035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,6144,2048,0.02944177720281813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,64,0.04852799905671013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,51200,0.15491910775502524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,32,0.005201777650250329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,51200,0.12046933174133301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,3584,0.00851999968290329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,64,0.013533333109484779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,10240,0.015917332635985482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,512,0.00518133325709237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,4096,0.00833511104186376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,51200,0.5674657821655273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,1536,0.03351377778583103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,12288,0.06128178040186564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,6144,0.19243821832868788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,2560,0.045112000571356885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,10240,0.03159555461671617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,128,0.00442755553457472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,3584,0.009391999906963771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3584,5120,0.01815022197034624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,6144,0.16567911042107478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,4096,0.019268444842762418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,3584,0.06313600142796834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1536,6144,0.01626044511795044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,1536,0.015078221758206686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,7168,0.02864444586965773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,128,0.0052915554907586836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,1024,0.14650221665700278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,5120,0.028220444917678833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,256,0.006981333096822103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,64,0.004139555411206352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,65536,0.10721777545081244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,12288,0.014952888091405233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,1536,0.6965449121263293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,16384,0.15889155864715576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,2048,0.009078222016493479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,65536,0.26539111137390137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,51200,0.07403644588258532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,512,0.009568000005351173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,128,0.007786666353543599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,6144,0.03671911027696397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,1536,0.016175111134847004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,6144,0.1422533326678806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,8192,0.09473333093855117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,1536,0.006221333311663733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,8192,0.024949333733982507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,7168,1.0286693572998047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,16384,0.044212443961037524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,7168,0.017895110779338412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,128,0.003556444413132138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,65536,0.2013031111823188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,10240,1536,0.09425599707497491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,2560,0.010527999864684211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,1024,0.005883555445406172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,51200,0.6247022416856554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,768,0.0064897777305708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,7168,0.034882667991850115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,8192,0.6679866578843859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,7168,0.09539466434054905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,65536,0.012773333324326409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,3072,0.00637511122557852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,64,0.006022222340106964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,1536,0.02202400068442027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,1536,0.006142222219043308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,1536,0.04526933365397983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,51200,0.40348267555236816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,256,0.004942222187916438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,128,2048,0.007113777928882175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,2560,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,16384,0.05135111014048258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,5120,0.032087110810809664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,3072,0.01643200053109063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,2560,0.03449511196878221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,256,0.005525333185990651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,5120,0.10045333041085137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,10240,0.040940443674723305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,10240,0.25313954883151585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,8192,0.018775999546051025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,12288,0.023357333408461675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,2048,0.11790488825903998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,2560,0.007909333540333642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,3584,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,16384,0.3342728879716661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,65536,0.4393831094106038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,51200,0.09526399771372478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,10240,0.02858577834235297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,16384,0.014366222752465142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,51200,2.9104906717936196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,3584,0.01854044364558326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,10240,1.318486213684082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,2048,0.015831111205948722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,10240,0.1838302214940389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,6144,0.031340443425708346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,16384,0.23509510358174643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,51200,0.08681777450773452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,128,0.004287111262480418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,512,0.010399111443095738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,1024,0.010832000109884473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,2048,0.01587733295228746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,32,0.007649777664078607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,6144,0.030745777818891738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,3584,0.06663999954859416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,2560,0.007000000112586551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,4096,512,0.10962755812538995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,12288,0.14952977498372397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,51200,0.46784088346693253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,1024,0.005621333503060871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,16384,0.032560000816980995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,8192,0.05613422393798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,8192,0.012502222425407834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,1536,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,10240,0.01902755598227183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,3584,0.007937777373525832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,65536,1.1346480051676433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,256,0.004460444466935264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,3584,0.023018666439586218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,512,0.010715555813577441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,768,0.01165688865714603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,512,0.009665778113736046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,4096,0.11698577139112686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,5120,0.177818669213189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,512,0.004933333231343163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,2560,0.0075528886583116316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,2560,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,10240,0.478290663825141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,16384,0.2433422141604953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,64,0.0069466663731469055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,64,0.005103111267089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,3584,0.11271733707851833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,16384,0.5230239762200249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,16384,0.23914665646023223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,4096,3072,0.014137778017255994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,3072,0.03307377629809909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,2048,0.13373243808746338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,51200,0.04831377665201823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,1024,0.011853333148691388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,4096,1.8002107408311632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,256,0.007913777397738563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,4096,0.01896711190541585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,12288,0.11315733856625027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,16384,0.05321955680847168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,12288,6.285377078586155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,512,0.018625777628686693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,1536,0.10011377599504258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,2048,0.009182222187519073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,65536,0.09097244342168172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,768,0.004879111217127906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,7168,0.23094489839341906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,16384,0.03010222315788269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,3072,0.0068959999415609576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,768,0.0038408889538711975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,2560,0.025412445267041523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,1024,0.02436977790461646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,4096,0.36223554611206055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,16384,0.023365333676338196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,65536,0.021574222379260596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,6144,0.014191110928853353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,768,0.01309866706530253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,2048,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,768,0.0058222222659322954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,5120,0.01055555542310079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,32,0.43663999769422746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,51200,0.09437866343392266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,8192,0.023952000670962866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,1024,0.0063786668082078295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,64,0.0073626670572492815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,7168,0.073999113506741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,512,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,3072,0.06635110908084445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,10240,0.013129777378506131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,512,0.006970666348934174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,16384,0.48163641823662656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,16384,0.08496799733903672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,2560,0.052684442864524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,64,0.038535111480289035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,2048,0.07726933558781941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,2048,0.2108853393130832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,3072,0.026335999369621277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,5120,0.11404089132944743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,32,0.005155555489990446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1536,1024,0.00564177789621883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,8192,0.01349422252840466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,2048,0.010632889138327705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,16384,0.1734639936023288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,32,0.05189955565664503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,65536,1.2260017395019531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,1024,0.004776000148720211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,1536,0.005707555760939916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,10240,0.05164711011780632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,768,0.015097777048746744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,2048,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,2048,0.05144266618622673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,6144,0.010860444770918952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,8192,0.24599022335476348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,5120,0.007121777368916406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,12288,0.12142133712768555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,1024,0.004871110949251386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,2048,0.0083137775460879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,3072,2.2400364345974393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,64,0.006140444427728653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,7168,0.46537690692477757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,8192,0.02238133384121789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,5120,0.0376160012351142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,16384,0.10124978091981675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,2048,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,2560,0.9024488661024305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,65536,0.5499342282613119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,32,0.006646222124497096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,768,51200,0.34393244319491917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,16384,0.07518044445249769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,32,0.002946666752298673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,7168,0.057429333527882896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,6144,0.03764266769091288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,768,0.036257776949140764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,1024,0.024489776955710515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,128,2560,0.006416888700591193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,51200,1024,0.2129502163992988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,51200,0.05305244525273641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,3072,0.10900355709923638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,1024,0.02446488870514764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,51200,1.341391987270779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,64,0.016581333345837064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,8192,0.5892426702711318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,4096,0.03964799973699782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,65536,2.9395262400309243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,2048,0.18648266792297363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,7168,0.2264782322777642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,4096,0.169432004292806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,64,0.007634667058785756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,256,0.010448888772063784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,5120,0.0849271085527208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,3584,0.01404355631934272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,10240,0.019291554888089497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,10240,0.04083289040459527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,512,7168,0.026678222748968337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,64,0.004320000194840961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,12288,0.025119110941886902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,512,0.006215999937719769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,32,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,16384,0.007678222325113084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,2560,0.007789333661397298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,5120,0.022504000200165644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,1024,0.023372444841596816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,1536,0.005474666754404704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,2048,0.015110222829712762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,64,0.005040888984998067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,10240,1.2091493606567383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,2048,0.10048977533976237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,10240,0.3137582143147786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,128,0.03458844290839301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,4096,0.007322666545708974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,2560,0.006053333481152852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,65536,1.580826653374566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,1536,0.01200177768866221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,512,0.009570666485362584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,1024,0.01499644418557485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,12288,1.1272186703152127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,4096,0.021048888564109802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,5120,0.1315120061238607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,768,0.04173511266708374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,10240,0.028696888022952612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,12288,0.3635271125369602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,3584,0.012657778130637275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,2560,0.8180604510837131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,7168,0.03614133265283372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,768,0.07132977909511991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,6144,0.31482844882541233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,51200,2.196407953898112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,2560,0.01815555493036906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,128,0.024078221784697637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,5120,0.0074240002367231585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,6144,0.02102577851878272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,3072,0.014390221900410123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,2560,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,4096,0.045238223340776235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,256,0.003107555624511507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,2048,0.007356444166766272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,32,0.01882577770286136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,768,0.012712000144852532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,8192,0.01276622215906779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,8192,0.2967031002044678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,12288,0.015828443898095023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,4096,0.008473777936564552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,512,0.016159110599093966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,8192,0.011917333636018964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,6144,0.018304889400800068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,65536,0.07476800017886691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,10240,4.660580529106988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,3584,0.060712887181176074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,3584,0.18478755156199136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,7168,0.015071999695565967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,2560,0.010138666464222802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,3584,0.02130577796035343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,4096,0.013891556196742587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,6144,0.050584889120525785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,3072,0.02602311140961117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,128,0.004113777644104428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,5120,0.012867555850081973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,51200,0.5661137898763021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,10240,4.773209889729817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,65536,0.8881120151943631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,1536,0.058740443653530545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,65536,2.231557422214084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,4096,0.14952622519599065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,64,0.006794666664467917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,256,0.006036444256703059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,7168,9.944906446668837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,768,0.02312444481584761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,3072,0.010038221875826517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,2560,0.09080977572335137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,256,0.006143110907740063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,768,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,512,0.018831110662884183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,51200,0.10891199774212307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,4096,5120,0.15530933274163142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,12288,0.06670222017500135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,12288,0.06322844160927667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,12288,2.6126391092936196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,4096,0.020219556159443326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,256,0.004393777913517422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,6144,0.01497244503762987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,256,0.030956443813112047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,5120,0.17326311270395914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,6144,0.02014844450685713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,3072,0.010997333460383944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,128,0.03753066725201077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,16384,2.657655927870009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,4096,0.24257956610785592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,2048,0.04296711087226868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,256,0.008408889174461365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,256,0.005195555587609609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,12288,0.018799111247062683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,512,0.006148444281684027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,3072,0.4685315555996365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,2560,0.017849778135617573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,6144,0.012180444267061023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,3072,0.12530222203996447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,10240,0.01683999929163191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,8192,0.0251511103577084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,32,0.005290666802061929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,64,0.0025884444928831523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,65536,2.687148412068685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,2048,0.022862222459581163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,7168,0.018452444010310702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,2560,0.16117955578698054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,768,0.03932355509863959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,64,0.0035208890007601846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,10240,0.06232800086339315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,7168,3584,0.06996266709433661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,3072,0.029261334074868098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,7168,0.10259377956390381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,16384,0.5658213297526041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,51200,0.42780354287889266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,2048,0.076200889216529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,256,0.00313688897424274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,6144,0.017840888765123155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,10240,32,0.006085333310895496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,12288,0.5813564194573296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,65536,0.5066453086005317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,3584,0.011165333290894827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,4096,0.008264888491895463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,1024,0.004916444420814514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,128,0.006503110958470239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,768,0.006116444451941385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,3072,0.1778115563922458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,7168,0.013245333400037555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,3072,0.009211555951171452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,32,0.003933333274390963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,3584,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,4096,0.032356444332334734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,7168,0.010962666736708747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,10240,0.016903999778959487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,256,0.010035555395815108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,8192,0.03015111221207513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,768,0.2673475477430555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,1024,0.021928888228204515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,1536,0.0343137780825297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,16384,1.3740195168389215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,768,0.008104888929261101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,16384,1.3528835508558485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,12288,0.008364444805516137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,2560,0.2632826699150933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,64,0.0159599996275372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,51200,0.03644355469279819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,3072,0.010017777482668558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,65536,0.8811848958333334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,1024,0.006065777606434292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,32,0.0053182223604785065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,2048,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,51200,0.7793386777242025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,5120,0.039375109805001154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,7168,0.249492433336046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,32,0.004953777624501122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,7168,0.2546782228681776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,256,0.04321066538492838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,7168,0.022329777479171753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,5120,0.022255111071798537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,128,0.008476444416575963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,16384,0.07208000289069282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,6144,0.012705778082211813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,8192,0.5091510878668891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,128,0.008307555483447181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,2048,0.0140328887436125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,1024,0.015113777584499784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,8192,0.019900444481107924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,10240,0.031103110975689355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,8192,0.02644088864326477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,3072,0.016467556357383728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,64,0.002518222149875429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,1536,0.0389395554860433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,768,0.01254666679435306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,65536,5.9490966796875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,10240,0.02071822186311086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,6144,0.024136889312002394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,128,0.04278044568167793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,256,0.005619555711746216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,512,0.006416000011894438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,128,0.005483555710977978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,1536,0.14041244983673096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,5120,0.030312889152103003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,12288,3.4044168260362415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,1024,0.15821066167619494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,32,0.003624000069167879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,1536,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,12288,0.0841582218805949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,512,0.04179199867778354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,6144,0.2682586775885688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,7168,0.7348595725165473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,65536,1.960102293226454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,256,0.01574133336544037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,1024,0.07488177882300483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,4096,0.10499999920527141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,2560,0.015632889337009855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,1024,0.01056977775361803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,12288,0.02676444417900509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,7168,0.013762666947311826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,65536,0.07035377952787611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,2560,0.011191111471917896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,16384,0.9579706192016602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,51200,0.06037066380182902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,1024,0.006976889239417181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,4096,12288,0.45367733637491864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,2560,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,1024,0.007197333706749811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,256,0.02937244375546773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,64,0.0031715554909573663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,768,0.008035555481910706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,4096,0.016806221670574613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,8192,0.023762666516833838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,6144,0.05985155370500353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,1536,0.00756888919406467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,5120,0.009932444327407414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,768,0.0324942237801022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,1536,0.005013333426581489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,256,0.003528888854715559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,128,0.060405333836873375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,3072,0.019165333774354722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,51200,0.23190045356750488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,2048,0.02163644466135237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,2560,0.08987644645902847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,64,64,0.00263288886182838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,5120,0.02552533315287696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,3072,0.007476444045702617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,12288,0.006600889066855113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,3072,0.0704666641023424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,2048,0.8403271039326986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,1536,0.011065777805116443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,3072,0.012366222010718452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,64,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,512,0.007629333270920648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,2048,0.01590844492117564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,256,0.004097777936193678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,256,0.003937777959638172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,51200,4096,0.10896799961725871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,65536,0.10828444692823623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,1024,0.03020355436537001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,2048,0.026894221703211468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,6144,0.022687110635969374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,6144,0.47274043824937606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,7168,0.2419511212242974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,6144,0.2388693226708306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,3072,0.009143999881214565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,10240,0.022858666049109563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,7168,0.056644446320003934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,128,256,0.005375999957323074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,768,0.00406666679514779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,7168,0.011012444065676795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,256,0.04012444284227159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,64,0.0026391111314296722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,2560,0.3975377877553304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,4096,0.02233244478702545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,512,0.008200888832410177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,4096,0.09061155716578166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,65536,0.07467555337482028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,51200,2.427439159817166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,10240,2048,0.08295822143554688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,3072,0.00832444429397583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,1536,0.07164355781343248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,1024,0.05951378080579969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,512,0.006457777900828256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,6144,0.03225866622394986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,2560,0.21414577960968018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,51200,0.12415466043684219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,32,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,768,0.004521777646409141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,10240,0.03551200032234192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,4096,0.010676444404655032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,128,0.013391111459996967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,4096,51200,0.5210880173577203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,12288,0.05197511116663615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,64,0.0046471113132105935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,65536,0.16035289234585232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,51200,2.2313866085476346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,3072,0.007517333659860823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,2560,0.010638222098350525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,16384,0.6151608890957302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,3072,0.14253777927822536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,256,0.005393777870469623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,64,0.01347644461525811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,5120,0.2908862166934543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,64,0.009452444811662039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,64,0.011172444456153445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,2560,0.017300445172521803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3072,51200,0.1951786677042643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,1024,0.013341333303186627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,2048,0.01108977778090371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,4096,8192,0.021946666969193354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,2560,0.026155556241671245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,64,0.0025617778301239014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,768,0.009538667069541084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,1536,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,64,0.011216888825098673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,32,0.0025537777692079544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,2048,0.008919999831252629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,2560,0.00703111125363244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,65536,0.4250097804599338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,128,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,5120,0.25560087627834743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,1024,0.725388420952691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,51200,0.16119910611046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,10240,0.02198222279548645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,12288,0.022520888182852004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,65536,0.35111112064785427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,256,0.05293066634072197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,10240,0.02037244372897678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,768,0.00720888872941335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,3584,0.1784728897942437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,2560,0.006796444455782573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,7168,0.10795911153157552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,6144,0.006174222048785951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,512,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,65536,9.340439690483942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,7168,9.661282857259115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,10240,0.1255626678466797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,256,0.020131554868486192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,64,0.002560888934466574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,2560,0.22894043392605254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,12288,0.02368533280160692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,8192,1.7752621968587239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,32,0.007626666790909237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,2560,0.05163733495606316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,6144,2.5293074713812933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,51200,0.13553688261244032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,32,0.009676444033781687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,12288,0.01884711119863722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,65536,0.1225386725531684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,51200,0.01146311147345437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,7168,0.10617155498928493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,8192,0.019150222341219585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,768,0.014381332529915703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,32,0.19719199339548746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,3072,0.013291555146376291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,64,0.006881777611043718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,768,0.021718222233984206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,8192,0.03715822100639343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,5120,0.011754666765530905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,32,0.0074986666440963745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,6144,0.024878222081396315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,32,0.002428444516327646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,2560,0.11504622300465901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,3072,0.04721422327889336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,3072,0.23089154561360678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,1024,0.02753511071205139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,5120,0.006585777633719974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,8192,0.013433777623706393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,5120,0.2890755600399441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,512,0.06914933522542317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,1024,0.08877688646316528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,12288,0.0464897784921858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1024,256,0.004530666602982415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,4096,0.014175110393100314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,8192,1.625485314263238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,512,0.005378666851255629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,4096,0.04584799872504341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,51200,0.043559110826916166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,7168,0.08612800306744045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,2048,0.09256888760460748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,5120,0.016399110356966656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,2560,0.014670222997665405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,65536,0.05913066864013672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,768,0.0803040001127455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,1536,0.009715555442704095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,12288,64,0.007491555478837755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,2560,0.007791111038790808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,4096,0.14497421847449407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,512,0.11483911673227946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,2048,0.05623288949330648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,1536,0.04092622134420607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,3072,0.10651289092169867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,5120,64,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,3584,0.15720800558725992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,12288,0.1424915525648329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,7168,0.011400000088744693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,7168,0.023037332627508376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,16384,2560,0.018744889232847426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,16384,0.06066399812698364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,16384,0.025429333249727886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,4096,0.1479840013715956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,16384,0.09185155232747395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,8192,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1536,51200,1.8010372585720484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,12288,0.10865333345201279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,5120,0.011368888947698804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,768,0.07695021894243029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,8192,0.03359377715322707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,2048,0.09846666786405776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,51200,0.6384862263997396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,51200,0.09999733501010472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,768,0.07928888665305243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,3072,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,256,0.004632888982693355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,10240,0.1258853276570638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,12288,0.040404445595211454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,51200,4.7845378451877165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,2048,0.007190222541491191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,1024,0.11121511459350586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,7168,51200,0.1362560060289171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,8192,0.3881857660081651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,12288,0.25570665465460884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,512,0.006907555378145642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,8192,0.006879999819729064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,4096,0.21059644222259521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3072,512,0.014472888575659858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,5120,0.39908266067504883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,7168,0.02017244365480211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,51200,0.06413155794143677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,64,0.006996444529957241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,12288,0.02200444373819563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,51200,0.01274400038851632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,1024,0.018453333112928603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,128,0.0033315554675128725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,1536,0.8980684280395508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,51200,0.10333333412806193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,128,0.0618106656604343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,10240,3.6381153530544705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,64,0.010584000084135266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,6144,0.017791999710930716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,128,0.031650665733549334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,1536,0.01255022237698237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,51200,1.5232008828057184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,65536,10.917693244086372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,6144,0.02064622276359134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,256,0.24629688262939453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,51200,12288,0.46907557381523984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,5120,1.2111262215508354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,51200,10.169605678982204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,6144,0.011159111228254108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,4096,0.017308443784713745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,65536,0.39805155330234104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,128,0.0064284445510970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,32,1024,0.010317333042621613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,3072,0.14835200044843885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,32,0.0032702222880389956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,51200,0.01644533375898997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,5120,0.02313333253065745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,7168,0.021724444296624925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2048,4096,0.018944000204404194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,512,0.04498399959670173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,16384,0.6637848748101128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,128,0.010576888918876648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,4096,0.023396443989541795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,2560,0.011824000212881299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,2560,0.041489776637819074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,128,0.00574222207069397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,3584,0.8895573086208768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,256,0.012854222622182635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,16384,0.02384533319208357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,6144,1.173199971516927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,10240,0.07970222499635485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,4096,0.00756177802880605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,512,0.037765334049860634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,128,0.005991999887757831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,128,0.007335111498832703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,4096,0.01277955538696713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,2048,0.19724622037675646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,4096,0.06387555599212646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,16384,0.038070221741994224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,3072,0.018613333503405254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,32,0.00281866660548581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,5120,0.17813777923583984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,12288,0.20766133732265898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,1024,0.0038319999972979226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,10240,0.013787555197874704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,768,0.008269333177142674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,65536,0.06538933515548706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,16384,0.07675200038486056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,2048,0.005538666827811136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,768,0.1180444426006741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,4096,0.0088746663596895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,16384,0.5328008863661025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,7168,0.010781333678298526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,51200,6144,22.855266994900173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,256,0.014396443963050842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,256,0.00848355558183458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,512,0.011632888681358762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,7168,0.028470221492979262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,8192,0.08293155829111735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,12288,0.03512888815667894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,128,0.007986666427718269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,10240,0.027599111199378967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,256,0.004948444250557157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,6144,0.05676711267895169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,4096,0.15466754966311985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,65536,0.20251911216311982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,12288,0.09467911058002049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,3072,0.10504088799158733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2560,8192,0.01809955636660258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,1024,0.0064604443808396654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,12288,0.021595555875036452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,2048,0.01924888955222236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,12288,0.21454044183095297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,256,0.007420444654093848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,1024,0.006649777707126405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,65536,0.6172480053371853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,32,0.0032222221295038858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,10240,0.037343111303117536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,3584,0.06878044207890828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,512,0.03775555557674832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,12288,0.040236443281173706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,128,0.007866666548781926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,16384,0.18542844719356963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,8192,0.2195671134524875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,2560,0.025557332568698462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,32,0.003948444293604957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,1024,0.01014311114947001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,128,0.004294222427739037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,64,0.012877333495352002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,32,0.008655110994974772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,12288,0.018014222383499146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,4096,0.03547644284036424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,256,0.013391111459996967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,768,0.005636444522274866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,8192,0.36288266711764866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,2560,0.012233778006500669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,6144,0.514844470553928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,1024,0.011503111157152386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,64,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,3584,0.010253333383136326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,1536,0.005896889087226655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,1536,0.01850488947497474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,128,0.0035137778355015647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,4096,0.017656000124083627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,16384,0.012120889292822944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,3584,0.009096889032257928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,10240,0.014866666661368476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,2560,0.007255111303594377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,5120,0.008917333351241218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,3584,1.1814026302761502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,3072,0.011712888876597086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,5120,0.02700800034734938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,5120,0.11941511101192898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,7168,0.01516533394654592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,65536,0.07044977611965604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,3584,0.007991111112965478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,2560,0.010595555934641095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2048,65536,1.0271422068277996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,3584,0.03139644530084398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,128,0.004898666507667965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,65536,0.6898951000637479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,5120,12288,0.040934221612082586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,256,0.018810666269726224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,5120,2.097225824991862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,2048,0.4595075713263617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,4096,0.9337493048773872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,1024,0.01092266705301073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,5120,0.018718222777048748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,1536,0.01550044450494978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,8192,0.01998133295112186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,51200,0.06684888733757867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,16384,108.04905870225694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1536,6144,0.4728613429599338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,768,0.026499556170569524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,5120,16384,0.037845333417256675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,51200,0.01868088874551985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,8192,0.6377609041002061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,1024,0.008228444390826756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,65536,0.08650133344862197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2560,4096,0.09425955348544651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,1536,0.006788444601827198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,256,0.011482666763994427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,32,0.0057546666098965546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,4096,0.16648622353871664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,5120,0.26205600632561576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,8192,0.16051822238498265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,12288,0.25071999761793345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,32,0.01184977756606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,1536,0.09600889020495945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,32,0.011659555964999728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,8192,0.01869333287080129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,512,0.1859288877911038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,12288,4096,0.022450667288568284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,10240,0.017447999782032438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,128,16384,0.05772888660430908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,1536,0.014730667074521383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,16384,0.04825244347254435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,7168,0.2330568896399604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,3072,0.04210222098562452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,5120,0.021427555216683283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,768,51200,0.7285626729329427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,4096,0.1288479963938395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,512,0.018619555566045973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,768,0.005694222119119432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,128,0.006158222340875202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,3584,0.028568890359666612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,1536,0.004794666750563515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,5120,2.032332526312934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,64,0.01295555548535453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,2560,0.015460444821251763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,32,0.01275644451379776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,5120,0.03522400061289469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,12288,0.09894844558503892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,1024,0.024392889605628118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,512,6144,0.05633866786956787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,51200,1.7727706697252061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,1024,0.0084915558497111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,8192,0.03594577643606398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,16384,5.129077487521701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,3584,0.3533919917212592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,3072,0.00998044427898195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,3072,0.010657777388890585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,1024,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,128,0.004879999905824661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,64,0.011132444772455426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,1536,0.02869866622818841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,512,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,16384,0.13130666149987116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,128,0.03383466601371765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,512,0.003964444415436851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,32,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,2048,0.04006933172543844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,2048,0.012836444709036084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,10240,1.1456107033623588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,10240,0.12909332911173502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,6144,1.768941349453396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,256,0.0805600020620558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,8192,0.14175466696421304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,512,0.003699555579158995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,256,0.007896000312434303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,5120,0.14712533685896131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,3072,0.009316444396972656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,16384,0.026016889346970454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,3072,0.014767110347747803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,51200,0.16925511095258924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,12288,0.017306667235162523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,2048,0.03262844350602891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,16384,0.1738791068394979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,64,0.01092088884777493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,256,0.0031644445326593188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,7168,0.017978666557206046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,16384,0.6779608726501465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,32,0.015188443991872998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,7168,0.09491111172570123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,2048,0.0073831114504072405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,5120,0.014274666706720987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,6144,0.08662222491370307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,5120,0.1772542264726427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,256,0.0038355555799272326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,16384,0.061569776799943715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,65536,0.35028799374898273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,8192,0.9156604342990451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,3072,0.010079999764760336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,32,0.004517333375083076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,768,0.012625777886973487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,3584,0.04199111130502489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,128,0.017883555756674874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,768,0.03251911203066508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,8192,0.030603554513719346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,768,0.05327466792530484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,3072,0.025197333759731714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,65536,0.0778684417406718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,10240,0.01312444441848331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,1024,0.019076445036464267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,65536,0.05694755580690172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,64,0.0031386667655573953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,7168,0.01108088923825158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,128,0.002801777794957161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,2048,0.006704888823959563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,128,0.013608888619475894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,12288,0.1041333344247606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,768,0.004028444488843282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,4096,0.015993777248594496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,1536,0.11715466446346706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,512,0.0058204444746176405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,4096,0.07413866784837511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,2560,1.4889163970947266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,2048,0.014408888088332282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,8192,0.5466088718838161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,16384,0.08742311265733506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,2560,0.0140346669488483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,10240,0.023997333314683702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,16384,0.9989226659138998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,5120,0.038246221012539335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,5120,0.014207111464606391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,65536,26.110241360134548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,2048,0.05694488684336344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,65536,4.06291029188368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,10240,0.041823112302356295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,2560,0.05766044722663032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,2560,0.008100444244013892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,4096,0.013836444252067141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,3072,0.013990221752060784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,2560,0.07648533582687378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,2048,0.005691555639108022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,16384,0.16304266452789307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,3584,0.010488889283604093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,10240,2.664327197604709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,3072,0.01405066748460134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,512,0.01351999988158544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,1024,0.007350222104125553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,768,0.005311111195219888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,10240,0.04477333360248142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,64,0.0627324448691474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,3584,0.01501066651609209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,8192,0.012624888784355588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,12288,1.5475982030232747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,512,0.2841911051008436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,768,0.010256888965765635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,256,0.1792808903588189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,10240,0.02120622164673275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,32,0.19010399447547066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,10240,0.019179556104871962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,2560,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,8192,0.026272000537978277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,51200,0.03382044368320041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,6144,0.043319112724728055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,2048,0.07091910971535577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,512,0.011858666936556498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,2560,0.009911999934249455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,2048,0.13901778062184653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,3072,0.10805422067642212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,16384,0.009696889254781935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,32,0.005361777626805835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,256,0.022408000297016565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,3584,0.009389333426952362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,12288,0.040575110250049166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,3584,0.010650667051474253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,3072,0.03387644555833604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,16384,0.3328942192925347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,65536,0.24928622775607637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,1536,0.02516888909869724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,3072,0.013397333522637686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,2560,0.5084284676445855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,16384,0.01940444442960951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,2560,0.044666667779286705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,6144,0.1664595603942871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,5120,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,51200,0.014831110835075378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,3584,0.010524444282054901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,12288,0.023123555713229712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,16384,10240,0.08330044481489393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,51200,3.3230169084337025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,8192,0.017870222528775532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,5120,0.02681422233581543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,1536,0.006870222174459034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,32,0.00508355556262864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,3072,0.14706667264302573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,12288,0.462831974029541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,1024,0.09856800238291423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,32,0.01571911076704661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,16384,0.07077599896325006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,2560,0.08689333332909478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,128,0.02782755427890354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,2048,0.058168888092041016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,2560,0.029779556724760268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,2560,0.01272266689274046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,128,0.005119110975000593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,51200,21.015506320529514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,64,0.022241777843899194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,128,0.0047173334492577445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,7168,0.01239822225438224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,65536,1.3217804167005751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,1536,0.03355377912521362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,3072,0.031678222947650485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,64,0.07827822367350261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,32,8192,0.006640888750553131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,65536,0.0680968893898858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,10240,0.02165866725974613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,65536,0.20770843823750815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,10240,0.03101244568824768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,1024,0.011078221930397881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,6144,0.2718302143944634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,1024,0.017712000343534682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,512,256,0.0034684443639384378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,12288,4.448568979899089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,3584,0.018578666779730055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,128,0.019654222660594516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,64,0.019102222389645047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,1536,0.03815022110939026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,64,0.009104889300134447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,6144,0.09950488805770874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,3584,0.028394665982988145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,7168,0.02494844463136461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,7168,0.023800889650980633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,51200,0.6432879765828451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,32,0.005218666460778978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,32,0.003160888950030009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,4096,0.05808266666200426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,51200,5.146966298421224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,8192,0.02223555578125848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,4096,0.013248888982666863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,65536,51200,1.6381022135416667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,51200,0.06703644328647189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,64,0.006495111104514863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,16384,0.41248178482055664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,8192,0.017511111166742113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,51200,5.089230431450738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,16384,0.2528791162702772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,2560,0.3840151098039415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,3584,0.015220445063379077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,128,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,4096,0.07417955663469103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,7168,0.011711111499203576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,6144,0.01311288856797748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,256,0.002977777686383989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,128,0.05465066764089796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,1024,0.005099555684460534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,3072,0.00923733330435223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,5120,0.02100533412562476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,32,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,12288,0.17468533251020643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,12288,0.09448799822065566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,10240,0.2660106552971734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,3072,0.00908622228437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,3072,0.06853777832455106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,7168,0.022984888818528917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,65536,0.12142666180928548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,512,0.012711111042234631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,8192,0.008160889148712158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,2048,0.01679733395576477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,64,0.00702400008837382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,8192,0.5314897961086696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,3584,0.016064888901180692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,64,0.3610106574164496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,2560,0.06277510854932997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,5120,0.00902844468752543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,7168,0.0726720028453403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,65536,0.3628355662027995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,51200,0.02709688908523983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,128,0.003875555677546395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,8192,0.025427555044492085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,1536,0.06317333380381267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,2048,0.01590488851070404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,2048,0.01090755561987559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,1536,0.006050666587220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,2560,0.09324000279108684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,256,0.002903999967707528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,10240,0.023797333240509033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,768,4096,0.04399111204677158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,2560,0.011622221933470832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,768,0.005625777774386936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,2560,0.008377778033415476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,1024,0.005480000128348668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,128,0.004495999879307217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,1536,0.008897778060701158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,8192,0.21351999706692168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,4096,0.038831998904546104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,8192,0.21209600236680773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,12288,0.02827733423974779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,64,0.009242666264375051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,3584,0.00684799998998642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,256,0.003960000144110786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,16384,0.006044444524579578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,3584,0.011184000306659274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,512,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,10240,0.026332444614834253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,65536,2560,0.14586222171783447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,1024,0.07542400227652656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,128,0.003979555434650845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,7168,0.162228438589308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,65536,0.1372915506362915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,3584,0.01094133324093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,1536,0.006723555425802867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,7168,0.12952889336480036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,5120,0.03155022197299533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,65536,2.3549840715196395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,1536,0.018019555343521964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,768,0.021459556288189355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,128,0.006931555353932911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,6144,0.046572443511750966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,128,512,0.0038613333470291565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,6144,0.012518222133318583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,1024,0.005169777820507686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,5120,0.010627555350462595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,12288,0.10140622324413723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,8192,0.02277244461907281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,128,0.004192000048028098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,12288,0.3163013458251953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,10240,0.064374221695794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,12288,0.05937777625189888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,65536,0.11869333850012885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,16384,0.030071109533309937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,8192,3.1325999365912547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,256,0.00685511115524504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,7168,0.04630577895376417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,2048,0.0063795554969045855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,3072,0.0073315559162033936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,7168,0.0061039999127388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,2048,0.005144888742102517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,128,0.015222221612930298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,3072,0.016731555263201397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,2560,0.011433777709801992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,4096,0.1411235597398546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,10240,0.09010933505164252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,10240,0.6550710995992025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,512,0.007473777565691207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,7168,0.07942489120695326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,256,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,16384,0.021335111724005804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,51200,0.03103111187616984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,1536,0.012527999778588613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,4096,512,0.005650666852792104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,512,0.04169688953293694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,512,0.007280000381999546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,3072,0.009473777479595607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,768,0.0058453331391016645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,7168,1.5241929160224066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,5120,0.01274311128589842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,768,0.00740533322095871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,128,0.06147466765509712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,2048,0.011200000014570026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,1024,0.00647199981742435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,32,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,6144,0.015448000695970325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,65536,0.06870844629075792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,6144,0.03335111008750068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,65536,2.606170654296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,10240,0.060552888446384005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,7168,0.0166631109184689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,32,0.005056000004212062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,512,0.004839111119508743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,5120,0.012181333369678922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,128,0.003112000102798144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,1024,0.03161777721510993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,10240,0.20263911618126762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,512,0.00535377777285046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,3072,0.06092800034417046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,1536,0.010858666565683154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,768,0.004105777790149053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,2048,0.007150222029950883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,8192,0.01587022178702884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,16384,0.01017955525053872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,65536,1.0067893134223092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,256,0.005509333478079901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,1536,0.006151111175616582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,2048,0.05997955799102783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,51200,0.2793404526180691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,768,0.007961778177155389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,256,0.03271289004219903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,10240,0.05309066507551405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,32,0.016713778177897137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,256,0.0053075556125905775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,64,0.0028044444819291434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,10240,0.6622773276435004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,3072,0.12967644797431097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,1536,0.018771555688646104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,512,0.022125333547592163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,512,0.005393777870469623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,7168,2.786253399319119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,128,0.004034666551484002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,51200,0.6530444357130263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,4096,0.05122133427196079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,3072,1.1279208925035265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,512,0.0042035554846127825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,65536,0.41127999623616535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,64,0.003120888852410846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,128,0.006024000131421619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,5120,0.05456977751519945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,6144,0.022950222094853718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,2560,0.3706977897220188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,4096,0.054551111327277295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,1536,0.008649778034951951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,5120,0.09333688682980007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,3584,0.1997599999109904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,12288,0.021336000826623704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,16384,0.022403554783927068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,1024,0.0058559998869895935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,1536,0.01109688894616233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,51200,0.7414346800910102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,768,12288,0.017811555001470778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,65536,0.10537955496046279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,7168,1.506808916727702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,4096,0.023039999935362074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,51200,0.07335822449790107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,16384,0.028331554598278467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,6144,0.021497777766651575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,3584,0.17689777745140922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,32,10240,0.16043378247155085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,5120,0.43854665756225586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,65536,3584,1.7661333084106445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,1024,0.06229866875542534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,65536,0.3676151169670953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,128,0.004447999927732679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,4096,0.1105075544781155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,768,0.03126222226354811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,128,0.0058088890380329555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,12288,0.751989311642117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,6144,0.036370668146345354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,51200,1.4976906246609156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,768,0.017421333326233756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,6144,0.015325332681337992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,2560,0.009979555176364051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,4096,1.3447155422634547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,8192,0.05249333381652832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,8192,0.11592533853318955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,8192,0.031523555517196655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,51200,24.252939860026043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,51200,0.2995102140638563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,512,0.007045333584149678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,65536,0.2498577700720893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,6144,2.587879180908203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,10240,0.2964373429616292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,1536,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,5120,0.2538035445743137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,10240,0.024480000138282776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,64,0.003121777748068174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,1536,0.023702222439977858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,1024,0.011202666494581433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,128,0.012511110968059964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,64,0.014860444598727755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,65536,3.5332667032877603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,3072,0.012152888708644442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,6144,0.018466666340827942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,768,0.1410026682747735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,6144,0.027152889304690894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,512,10240,0.07527199718687269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,64,32,0.0033280000918441345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,2048,0.03647822141647339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,65536,0.5278373294406468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,3072,0.01516266663869222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,51200,2.4048898484971786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,7168,2560,0.022465777066018846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,512,0.005929777605666055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,256,0.006051555689838197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2048,16384,0.23240976863437227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,64,0.011210666762457954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,7168,0.2810888820224338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,51200,0.045178665055168994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,7168,2.23032357957628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,3072,0.011443555355072021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,7168,0.05232355660862393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,1536,0.005605333381228977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,1024,0.006483555667930179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,3584,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,3584,0.10493155320485432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,4096,1.460579554239909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,4096,12288,0.7502702077229818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,5120,0.01402311192618476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,1024,0.03440533412827386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,4096,0.017047999633683097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,1024,0.005738666488064661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,6144,0.16242399480607775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,128,0.010255999863147736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,51200,0.010660444696744284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,51200,1.4205146365695531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,5120,6144,0.01850133306450314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,16384,0.543923536936442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,128,0.03138666682773166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,5120,0.012947555217478009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,2048,0.008013333711359236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,12288,0.06551555792490642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,1024,0.009190222455395592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,256,0.028572443458769057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,7168,0.010909332997269101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,4096,0.008375999828179678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,16384,0.04822133315934075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,1024,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,3072,0.010767111347781287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,3072,0.008160889148712158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,256,0.02516799999607934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,4096,0.02587911155488756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,512,0.07021333111657037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,1536,0.011801777614487542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,2048,0.010638222098350525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,6144,5.414815266927083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,2048,0.15475200282202825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,6144,0.015062222878138224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,7168,0.03310399916436937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,12288,0.020633776982625324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,3584,0.03936800029542711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,3584,0.015665777855449252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,256,0.024531554844644334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,16384,0.12582755088806152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,5120,0.039113776551352605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,4096,0.013460444079505073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,3072,0.026597332623269822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,5120,0.011711999773979187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,512,0.005417777846256892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,512,0.006294222341643439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,51200,1.9038639068603516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,7168,0.03558933403756883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,6144,512,0.006061333335108227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,768,0.05405511127577888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,8192,0.04608266552289327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,8192,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,65536,1.967440923055013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,512,0.02292088833120134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,5120,0.05397955576578776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,10240,0.04821777674886915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,1024,0.010535111029942831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,3072,0.006884444504976273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,512,0.026200888885392085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,1024,0.013074666261672974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,12288,0.6654302279154459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,32,0.048751112487581044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,8192,0.01038577738735411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,3072,0.3810559908548991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,128,0.007352888584136963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,2560,0.0063635553750726916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,10240,0.5077155431111654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,3072,0.8050666915045844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,3072,0.067921777566274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,6144,0.029840889904234145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,51200,0.42828888363308376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,768,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,768,0.020429333051045735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,6144,0.0127688886390792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,4096,0.05199022094408671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,4096,0.03436444534195794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,64,0.005027555757098728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2048,64,0.0029528888149393927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,768,0.004284444368547863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,8192,0.029713776376512315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,4096,0.011615999870830111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,32,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,2048,0.01722133325205909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,128,0.005131555514203178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,256,0.003124444435040156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,1536,0.006559999866618051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,8192,0.0310204459561242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,2048,0.05309155583381653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,1024,0.005839111076460944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,5120,0.021010667085647583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,128,0.0033973333322339584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,65536,1.3691662682427301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,3072,0.019903111788961623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,128,0.057097779379950635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,6144,0.010327111515733931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,768,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,512,0.003432000055909157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,2048,0.27732621298895943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,12288,0.043763554758495755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,2560,0.011515555282433828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,7168,0.012366222010718452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,2048,0.008015111088752747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,7168,0.19395200411478677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,1024,0.005912888795137405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,768,0.009469333622190688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,64,0.0026088888860411118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,16384,0.03153866529464722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,6144,0.45618401633368594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,65536,0.1293155617184109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,256,0.028385778268178303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,8192,0.012703110774358114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,7168,0.2128711144129435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,10240,0.018651554981867473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,256,0.004474666797452503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,128,65536,0.07005777623918322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,8192,0.02589955594804552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,128,0.014123555686738757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,8192,0.0398231115606096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,8192,0.2850524319542779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,8192,0.013552889227867126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,5120,0.011074666347768573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,1536,0.015198222464985318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,2560,0.019467555814319186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,128,0.009473777479595607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,8192,0.13670667012532553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,2560,0.01818933255142636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,256,0.0038311113086011675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,8192,0.011735111474990845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,7168,0.06833689080344306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,51200,0.0788488917880588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,64,0.010198222266303169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,3072,0.014779556128713818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,3072,0.015603555573357476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,5120,1024,0.0070222218831380205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,768,10240,0.017633777525689866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,512,0.003552000141806073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,51200,0.2265333334604899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,7168,0.030268443955315485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,6144,0.01643111142847273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,2560,0.014782221780882942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,256,0.004228444563017951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,6144,0.009826666778988307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,256,0.05576444334454007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,64,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,2560,0.04025600022739834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,1024,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,6144,0.01536977787812551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,6144,2048,0.09090755383173625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,64,0.0028782222006056043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,12288,2.031902313232422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,3584,0.01221155540810691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,16384,0.09163555834028457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,51200,0.03258311086230808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,4096,0.008629333641793992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,6144,0.014334221680959066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,51200,0.1012515558136834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,4096,0.020270221763186984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,512,0.0033928888539473214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,768,0.032221333848105535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,1536,0.07447999715805054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,3072,0.011004444625642566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,7168,0.14180710580613878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,12288,0.07893333170149061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,4096,0.951258659362793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,512,0.021458667185571458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,1536,0.010830221904648675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,2560,0.9724986818101672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,32,0.002971555623743269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,2560,1.3441937764485676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,3072,0.09899733463923137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,256,0.006226666685607698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,10240,0.14062578148312038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,1024,0.005981333139869902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,3584,0.19943022727966309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,64,0.008591111335489485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,1536,0.03689688775274489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,768,0.22761866781446669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,51200,2.001174290974935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,5120,0.011795555551846823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,256,0.05933244360817803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,64,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,51200,0.17336089081234404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,7168,0.04508444335725573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,1024,0.006206222292449739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,2048,0.009425777528021071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3584,10240,0.09777155849668716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,8192,0.02826133370399475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,51200,0.09180355734295315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,2560,0.04487733377350701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,65536,0.25166132715013295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,256,0.0047351109484831495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,6144,0.010560000108347999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,10240,0.006903999795516332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,4096,0.8389324612087674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,4096,65536,0.14390933513641357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,10240,0.31348978148566353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,4096,0.011572444604502784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,7168,0.5125030941433376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,7168,0.02463022205564711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,3072,0.02086488902568817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,8192,0.0381991101635827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,512,51200,0.31667643123202854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,128,0.006224888894293044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,3072,0.6609662373860677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,10240,0.08890044689178467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,1024,0.013722666435771518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,3584,0.30785245365566677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,4096,0.04342666599485609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,3584,0.016187555260128446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,6144,0.09318844477335612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,5120,0.01700355609258016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,3584,0.018256889449225534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,4096,0.01789777808719211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,256,0.0084906667470932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,51200,0.36634845203823513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,12288,0.009103111094898647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,3584,0.08435733450783624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,128,0.008233778178691864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2048,65536,1.3639787038167317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,51200,0.9865635765923394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,4096,0.02927910950448778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,4096,0.011123555401961008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,65536,0.11604621675279404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,3584,0.009207111265924241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,5120,0.144195556640625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,51200,0.07416888740327623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,2048,0.0689200030432807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,128,0.0029377777957253983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,5120,0.7124222119649252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,512,0.004472889006137848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,4096,0.025363556212849084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,2560,0.01329066687160068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,768,0.004543111142185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,51200,0.14314044846428767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,6144,0.009168888959619734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,51200,0.19028355015648735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,65536,0.08152266343434651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1024,64,0.003248888792263137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,1536,0.035000890493392944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,10240,0.02474311159716712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,12288,0.026416889495319788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,768,0.010125333236323463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,2048,0.007039111521508958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,7168,0.025016888976097107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,512,0.005373333477311664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,128,0.003119999956753519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,2048,0.01145244472556644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,4096,0.017483555608325534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,2560,0.006584888945023219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,64,0.01517599986659156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,256,0.005614222337802251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,768,0.019183110859658983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,512,0.0074879998962084455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,65536,0.547377798292372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,51200,0.05452177921930949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,16384,0.11886044343312581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,128,0.004776000148720211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,6144,0.3104871114095052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,2048,0.0088764445649253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,7168,0.015081778168678284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,128,1536,0.0058257778485616045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,7168,1536,0.06548888815773858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,12288,0.5987084706624349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,1024,0.02846577763557434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,2048,0.006894222150246303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,10240,0.04365422328313192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,8192,0.236753781636556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,512,0.005523555394675996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,1024,0.004146666576464971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,512,0.009676444033781687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,512,0.00342399999499321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,4096,0.017910222212473553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,6144,0.020615999897321064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,32,0.008414222134484185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,64,0.016352888610627916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,2560,0.0069520001610120135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,7168,0.04036800066630045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,64,0.005256888767083486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,16384,0.8704613579644097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,128,0.01719199948840671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,4096,0.021477333373493616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,5120,0.42960532506306964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,4096,0.04900622367858887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,32,0.007460444337791867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,12288,0.017249777913093567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,256,0.014898666077189975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,8192,0.31928979025946724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,7168,0.524879985385471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,1536,0.006055111272467508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,10240,8.58512454562717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,1024,0.0064897777305708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,8192,0.02293599976433648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,4096,0.011870221959220039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,1024,0.010081777969996134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,3584,0.011743110915025076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,256,0.0031306667046414483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,3072,0.07803199688593547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,6144,0.01037688884470198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,3072,0.01621866722901662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,2048,0.020728000336223178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,51200,0.2450951205359565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,2560,0.0385155545340644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,8192,0.23064888848198783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,512,0.0052044445441828836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,64,0.003863999827040566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,1536,0.009453333086437648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,65536,5120,1.968017790052626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,65536,0.20456355147891572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,16384,0.557462215423584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,128,0.0028284444577164124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,3072,0.022897778285874262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,16384,0.024345777100986902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,7168,0.05262222223811679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,10240,7.70157453748915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,3584,0.01069777790043089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,3584,0.03867288761668735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,1536,0.010500444306267632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,4096,0.020821332931518555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,65536,0.11659910943773057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,10240,0.04576799935764737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,4096,0.2935262256198459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,12288,0.31795912318759495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,128,128,0.0031102223114834893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,1536,0.009450666606426239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,512,0.016505777835845947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,1536,0.056391113334231906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,512,0.019519999623298645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,1536,0.009672889278994666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,5120,0.18614844481150308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,1536,0.04271911250220405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,5120,0.013357333011097379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,768,0.007629333270920648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,5120,0.006866666591829724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,6144,64,0.018007111218240526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,3072,0.009221333596441481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,65536,512,0.27703022956848145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,256,0.007703999678293864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,7168,0.8521706793043348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,2560,0.19148266315460205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,3584,0.02165599995189243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,51200,0.19482666916317412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,6144,0.014375110467274984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,2560,0.051017777787314526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,12288,0.01899111105336083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,128,0.004927999857399198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,51200,0.056877334912618004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,12288,2.339303970336914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,768,0.012843555874294706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,768,0.012237333589129977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,7168,0.022307554880777996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,64,0.0026213334252436957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,3072,0.007419555551475949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,512,0.005009777843952179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,512,0.00847466703918245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,256,0.005142222262091107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,512,0.018722666634453666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,65536,0.07380977604124281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,10240,0.05666933457056681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,6144,0.021996445126003687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3072,768,0.005879111174080107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,1024,0.006011555592219035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,6144,0.014167111780908374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,2048,0.013815110756291283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,65536,143.08563910590277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,64,0.011794666449228922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,3072,0.006505777852402792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,16384,0.1405244403415256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3072,8192,0.16852266258663604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,1024,0.006272888845867581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,7168,0.06393955813513862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,51200,0.013663999736309052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,768,0.017129777206314933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,512,0.019880000087949965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,51200,0.08874577946133083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,6144,0.013772444592581855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,1536,0.052616000175476074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,3072,0.015396444333924187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,51200,0.056186669402652316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,512,0.005392000079154968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,10240,0.10599199930826823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,5120,0.015475556254386902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,128,0.05080622103479174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,768,0.008820444345474243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,8192,0.011066666907734342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,2048,0.009671111073758867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,16384,0.0502106679810418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,32,0.005207111024194294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,3072,0.13620355394151476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,7168,0.19207199414571127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,2048,0.024115555816226538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,64,0.0031031111462248694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,8192,51200,1.3517395655314128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,8192,0.058635552724202476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,12288,0.19225066237979463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,1536,0.004587555511130227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,1024,0.03790222273932563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,7168,0.08073688877953424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,3072,0.1518604490492079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,3072,0.027621333797772724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,6144,0.14769421683417425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,2560,0.02258488867017958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,10240,0.03257866700490316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,5120,0.32038844956292045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,16384,0.31099287668863934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,768,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,256,0.005162666655249066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,7168,0.33283999231126576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,2048,0.1788711150487264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,65536,1.5044292873806422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,64,0.0026151111556424033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,2560,0.33131111992730033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,65536,0.18924800554911295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,1024,0.18010400401221383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,4096,0.01798577772246467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,64,0.01459111107720269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,8192,0.00655377780397733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,6144,1.262296888563368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,2560,0.02089333368672265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,2048,1.0287351608276367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,10240,0.02929688824547662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,256,0.03564533260133531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,3584,0.11029066642125447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,16384,0.5665084520975748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,4096,0.05407822132110596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,8192,4096,0.01628533336851332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,64,0.0033484444850020935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,512,0.022589332527584497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,2560,0.05954044395022922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,1536,0.004576888763242298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,65536,0.9167066150241427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1024,51200,0.03501866592301263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,2560,0.9081652959187826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,512,0.006166222194830577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,2048,0.007999999655617608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,64,0.0059075554211934405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,1024,0.004970666848950916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,65536,0.3474586539798313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,6144,0.01703644461101956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,6144,0.3532488875918918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,1536,0.009720000127951304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,4096,0.016650666793187458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,10240,0.01760444376203749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,2048,0.009539555344316695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1536,4096,0.011516444385051727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,768,0.007898666792445714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,256,65536,0.021943999661339655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,5120,0.009686222506894005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,256,0.025029333101378545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,10240,0.05495822098520067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,51200,0.054213331805335156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,1536,0.030363556411531236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,7168,0.014911999305089315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,32,0.0028293333533737394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,51200,2.5681209564208984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,3584,0.014942222171359591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,2560,0.015683554940753512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,5120,0.1689200003941854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,51200,10.993278503417969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,3584,0.02442311081621382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,1536,0.06035555733574761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,10240,14.557957967122396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,5120,0.1518711116578844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,10240,0.6650222142537435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,2560,0.009222221871217092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,1536,0.01769333415561252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,5120,1.8100106981065538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,16384,0.09056622452206081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,128,0.003141333245568805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,5120,0.01762488815519545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,64,0.005967111223273807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,6144,0.018739556272824604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,3584,0.01239999963177575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,16384,0.11712177594502766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,5120,0.13061511516571045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,3072,0.0976053343878852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,32,0.6950737635294596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,65536,0.23727467324998644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,6144,0.11836355262332493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,8192,3.617473602294922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,768,0.0046746668716271715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,512,0.27635110749138725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,65536,0.4041582213507758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,2048,0.03314755691422357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,3584,0.03975288735495673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,768,0.3133297761281331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,1536,0.004454222404294544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,3072,0.010755555497275459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,5120,0.0380675560898251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,51200,0.13795999685923258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,10240,0.045144889089796275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,8192,0.026011554731263056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2048,256,0.0046746668716271715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,1024,0.041894220643573336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,1024,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,6144,0.03724088933732774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3072,51200,1.0584595998128254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,512,0.004087999877002505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,65536,0.09583110941780938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,51200,2.4372986687554254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,64,0.0063075555695427795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,6144,0.10821688837475246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,8192,2.45560540093316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,3072,0.1252568827735053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,16384,0.3394791020287408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,2560,0.11705866124894883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,8192,0.6090790960523818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,10240,0.048715555005603366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,8192,0.01458666721979777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,512,0.003415999934077263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,2048,0.04131022095680237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,8192,0.1659288936191135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,16384,0.02219911085234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,256,0.041351109743118286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,512,0.23280178176032174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,32,0.0027599999060233435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,1024,0.004252444538805219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,3584,0.025055999557177227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,8192,0.04069955481423272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,6144,0.4113315476311578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,512,0.015456888410780164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,32,0.034953776333067156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,512,0.04433511031998528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,3072,0.25854844517178005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,16384,0.2151528861787584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,1536,0.004931555440028508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,7168,0.023615999354256526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,1536,0.006260444306664997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,32,0.0037848887344201407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,3584,0.0233297778500451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,16384,0.4070088863372803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,5120,0.13539022869533962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,128,0.006292444550328785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3072,1536,0.01016533292002148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1024,5120,0.06801599926418729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,6144,0.10959466960695054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,5120,12288,0.4979199833340115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,256,0.006918222126033571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,10240,0.018611555298169453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,512,51200,0.17510933346218535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2048,256,0.03874311182234023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,12288,1.4645653830634222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,128,0.17349421977996826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,16384,0.2666257752312554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,3584,0.04127555423312717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,51200,2.872008853488498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,16384,0.9657511181301541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,3584,0.013952889376216464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2560,512,0.004559999952713649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,256,0.006089777582221561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,512,0.0041777777175108595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,1536,0.6180568801032172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,64,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,6144,0.01332266628742218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,32,0.002847111059559716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,512,0.007848000360859765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,3584,0.016113777955373127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,3584,0.016795555750528973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,6144,1.9690986209445531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1024,1536,0.008609777523411645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,65536,0.03877866599294875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,7168,0.018277333842383493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,32,0.006076444354322221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,16384,0.16172088517083064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,65536,0.07504977782567342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,256,0.026710222164789837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,1536,0.007327111230956183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,3584,0.27251021067301434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,768,0.010750222537252637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,1024,0.005720888988839255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,2048,0.01552977826860216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,65536,0.11933333343929714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,768,0.005685333162546158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,6144,0.009848888549539778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,6144,0.024134222004148696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,16384,0.00868088917599784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,64,0.010647999743620554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,16384,6144,0.031255112753974065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,3584,0.4923564592997233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,512,0.005599110904667113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,2048,0.17236622174580893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,10240,0.014594667487674289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,65536,3.9603394402398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,4096,0.013636444177892474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,1536,0.012248888611793518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,128,0.021643555826610986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,12288,0.01701244380739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,16384,0.028095109595192805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,256,0.005462222215202119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,10240,0.1506053341759576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,51200,0.18349688582950166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,1536,0.3771324422624376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,7168,0.0123857781291008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,256,0.03269688950644599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,3584,2.0120737287733292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,12288,0.5408479902479384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,7168,0.09789866871303982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,256,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,2560,0.008791999684439765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,16384,0.02036977807680766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,4096,0.013426666458447775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,768,0.0046737777690092725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,1536,0.008519111408127679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,4096,1024,0.01184177812602785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,32,0.0057946667075157166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,32,768,0.0033964444365766314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,768,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,2048,0.03355022271474203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,16384,0.4087031152513292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,512,0.003499555504984326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,128,0.017180444465743173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,6144,0.013071999781661563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,7168,0.1797564427057902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,4096,0.0075902218619982404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,4096,0.02716800073782603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,1024,0.019052444232834708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,128,0.011710222396585675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,3584,0.27926577462090385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,12288,0.890909300910102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,12288,0.050111999114354454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,512,0.04329866502020094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,16384,1.2147324879964192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,2048,0.021362667282422382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,32,0.04018755422698127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,7168,1536,0.02240444388654497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,6144,0.7158968713548449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,51200,0.5274826685587565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,512,0.0034560000316964257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,512,3584,0.006704000135262807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,65536,0.9871102439032661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,256,0.004369777937730153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,768,32,0.004373333520359463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,7168,0.006793777975771163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,3072,0.018018666240904067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,2048,0.029747555653254192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,256,0.005124444348944558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,4096,0.03804088963402642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,8192,0.02327377763059404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,12288,0.5603804588317871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,51200,1.9495608011881511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,6144,0.27221777704026967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,3072,0.4797164599100749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,3072,0.05572266711129082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,65536,4.382754855685763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,8192,0.1656764480802748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,768,0.005640888793600931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,2560,0.028378665447235107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,3072,2.8686141967773438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,5120,7168,0.08688800202475654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3584,2560,0.04155733187993368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,5120,0.01684711045689053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,128,0.039043555657068886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,512,0.8519057697719998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,3584,0.01960711181163788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,7168,0.04196177919705709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,4096,0.01995733380317688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,2560,0.027506666051016912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,16384,0.055925336149003774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,512,0.04181955589188469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,512,0.019725332657496136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,128,0.007461333440409766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,2560,0.007996444072988298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,7168,0.012186666329701742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,51200,0.40427732467651367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,7168,0.11157333850860596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,8192,0.01809511085351308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,51200,0.7170053588019477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,256,0.006065777606434292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,65536,0.30904444058736164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,3072,0.02419822249147627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,4096,0.3024844328562419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,2560,0.022128888302379187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,768,0.018176888426144917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,3072,0.16279289457533094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,16384,0.2927102247873942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,6144,1.1796809302435982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,5120,0.01456622282663981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,8192,8192,0.02907466557290819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,512,0.005690666536490123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,1536,0.035439110464519925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,32,0.004190222256713443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,32,0.006763555523422029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,512,0.009024889104896123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,7168,0.49000268512301975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,256,128,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,3584,0.8079644309149848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,16384,0.4311830997467041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,1536,0.03875289029545254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,2048,0.008444444172912175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,16384,0.09069422218534683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,6144,0.17311733298831514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,256,0.005467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,12288,0.1552657816145155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,16384,256,0.2489208910200331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,64,0.03986577855216132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,3584,0.011714666253990598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,6144,0.09515111313925849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,6144,0.010047111246320937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,512,0.0036435553597079385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,12288,0.04395111070738899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,768,0.014169777433077494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,64,8192,0.013915555344687568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,1536,0.02499377727508545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,2048,0.007840889195601145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1536,7168,0.0970275534523858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,12288,0.007062222394678328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,512,0.006010666489601135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,10240,0.01711466742886437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2048,768,0.017483555608325534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,1024,0.009639110830095079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,4096,0.9661049313015408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,512,0.12944177786509195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,128,0.0035955554081334006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,64,0.016522667474216886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,10240,0.08461155494054158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,32,0.005454222361246745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,128,0.04580266608132256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,8192,0.13054666254255506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,5120,0.030249777767393324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,4096,0.0384808878103892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,10240,0.2819235589769152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,6144,0.013339555925793119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,2560,0.007738666401969061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,256,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,6144,0.007221333682537079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,3072,0.016351110405392118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,10240,0.019671999745898776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,5120,0.22961065504286024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,768,0.006351999938488007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,65536,3584,5.944905598958333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,7168,0.04973333411746555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,10240,0.009544000029563904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,10240,0.024896888269318476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,8192,0.016207999653286405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,4096,0.0663599967956543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,64,0.0177875558535258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,128,0.010963555839326648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,256,0.0036888888312710654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,1024,0.005764444255166584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,12288,0.015272888872358533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,1024,0.023108444280094568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,5120,0.009273777405420939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,256,0.007038222418891058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,10240,0.05453688899676005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,1024,0.0058311112225055695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,3072,0.008692444198661381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,768,0.05861777729458279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,65536,4.088192833794488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,5120,0.0687555538283454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2560,256,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,1024,0.11204889085557725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,65536,10240,0.17923466364542642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,32,16384,0.030560887522167627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,7168,0.03299555513593886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,65536,0.3522373305426703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,128,0.008843555218643611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,32,0.018962666392326355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,7168,0.04183822207980686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,1536,0.06519733534918891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,1024,0.03141599893569946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,64,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,8192,1.5249386893378363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,6144,0.012578667038016848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,1024,0.012870222330093384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,3072,0.03009777930047777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,2560,0.057560000154707164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,128,0.004754666652944353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,2048,0.012330667012267642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,3584,0.017712000343534682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,2560,0.06526577472686768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,5120,0.026748443643252056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,16384,0.022727999422285292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,6144,512,0.00868888861603207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,16384,0.039649777942233615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,128,0.3382231129540338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,12288,0.040392001469930015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,5120,0.3618568844265408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,3584,0.15718400478363037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,4096,0.023020444644822016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,1536,0.006806222101052602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,6144,0.05769599808586968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,51200,0.057441777653164335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,6144,0.048560000128216214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,512,0.046421332491768726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,64,0.005479111025730769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,65536,0.47178930706448025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,16384,0.14204088846842447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,1536,0.01402933398882548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,3072,0.36554577615525985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,51200,0.16875378290812174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,768,0.02139644490347968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,10240,0.20646044943067762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,12288,0.5833804342481825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,6144,0.16387289100223118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,4096,0.07703377803166707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,512,0.01386844449573093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,5120,0.3439715438418918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,128,0.01622222198380364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,8192,0.017166222135225933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,32,0.006649777707126405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,16384,0.0531582236289978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,1536,0.005751111027267244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,5120,0.04878222280078464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,2048,0.004939555707905027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,7168,0.011004444625642566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,3072,0.021384888225131567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,3584,0.03680800067053901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,7168,0.028548445966508653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,512,0.03237244486808777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,8192,12288,0.058352000183529325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,2560,0.018915555543369718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,2048,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,2560,0.01162755572133594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,3072,0.014030221435758801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,16384,0.06006222301059299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,6144,5120,1.6741982565985785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,10240,0.04805422160360548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,7168,0.006602666858169768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,12288,6144,0.0365973346763187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,768,0.03571822245915731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,16384,0.007800000409285228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,2560,0.012507555385430654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,512,0.009415110780133141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,4096,0.01072799993885888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,2048,0.024243555135197107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,2048,0.07058933046129015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,128,0.004397333496146732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,256,0.004859555512666702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,512,0.038344889879226685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,5120,0.014730667074521383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,1024,0.007335111498832703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,32,0.011207111179828644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,5120,0.012255111502276527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,1024,0.019768888751665752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,1536,0.8461182382371691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,10240,0.037631111012564764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,7168,0.029570665624406602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,65536,0.33888178401523167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,768,0.006315555423498154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,1024,0.03310399916436937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,6144,0.011155555645624796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,128,0.03390311201413473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,10240,2.338372548421224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,65536,0.22278756565517852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,4096,0.14760533968607584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,6144,0.04327466752794054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,8192,0.012692444854312472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,6144,0.2344364325205485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,256,0.0069466663731469055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,5120,0.034259554412629865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,768,0.03453777896033393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,128,0.029817776547537908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,4096,0.2671484417385525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,12288,0.11435911390516494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,65536,0.3151822090148926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,3584,0.09034933646519978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,4096,0.01090666651725769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,4096,0.008455110920800103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,65536,1.7472302118937175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,5120,0.01368444412946701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,5120,0.012095111111799875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,4096,0.05836800071928236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,16384,1.1506008572048612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,16384,0.5129511091444228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,8192,0.17181244161393908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,12288,0.4378560119205051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,512,0.003560888684458203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,64,0.002963555562827322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,12288,0.10294044680065578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3072,128,0.015538665983412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,128,0.0041519999504089355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,7168,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,10240,0.022877333892716303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,2560,0.1379315588209364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,6144,0.16429244147406682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,768,0.003896000070704354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,4096,0.0625786648856269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,768,0.01092977739042706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,10240,0.013008889224794177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,5120,0.012881778180599213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,768,0.004280888785918553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,2048,0.16040178140004477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,2560,0.00849599970711602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,768,0.004941333499219683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,768,0.005161777966552311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,3584,0.05322844452328152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,51200,3.160377714369032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,1024,0.004828444371620814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,32,0.013548444542619916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,3584,0.008328888979223039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,1536,0.31304889255099827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,10240,0.024200000696712073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,3072,0.1613724496629503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,12288,0.12318578031327988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,16384,0.15840444299909803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,12288,0.4095128907097711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,3584,0.016325333052211337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,768,0.008740444150235917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,2560,0.02829866607983907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,5120,0.011605333122942181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,51200,0.18212710486518013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,4096,0.05536355574925741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,12288,0.12378044923146565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2048,3584,0.009530666801664565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,51200,0.13277688291337755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,3072,0.050233778026368886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,2048,0.06887822018729316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,12288,0.01869422197341919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,64,0.006605333338181178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,4096,0.05598400036493937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,65536,0.1778657833735148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,6144,0.0259226659933726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,1536,0.041377776198916964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,16384,0.36102133327060276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,12288,9.926082187228733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,256,0.016018667154841952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,16384,0.10939821932050918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,1536,0.006928000185224745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,12288,0.015660444895426433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,3584,0.007048000064161088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,8192,0.02052622205681271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,12288,0.3061182234022352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,512,0.03148355417781406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,768,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,1536,0.004716444346639845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,3584,0.006259555617968242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,7168,0.041175110472573176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,6144,0.2054000033272637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,32,0.0025946667624844448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,7168,65536,0.5994879934522841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,6144,0.2567404376135932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,768,0.04293333490689596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,5120,0.010461333725187512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,16384,0.19897688759697807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,768,0.00811644477976693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,5120,0.05409511261516147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,10240,0.03305777907371521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,4096,0.0551475551393297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,256,10240,0.006666666517655055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,1536,0.12393066618177627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,10240,0.06041422155168322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,256,0.03390222125583225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,32,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,256,0.1935573286480374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,1024,0.032905777295430504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,5120,0.11335199409061009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1536,2048,0.03820000092188517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,5120,0.020280000236299302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,51200,0.40522310468885636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,12288,0.26035910182529026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,2560,0.47615379757351345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,512,0.012649777862760754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,768,0.003583111282851961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,6144,0.016275554895401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,4096,0.006948444578382704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,3072,0.3438817660013835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,3072,0.10947022173139785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,3072,0.007672888537247975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,51200,0.22328265508015951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,65536,2.5870187547471786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,12288,0.03391999999682108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,512,0.0053137776752312975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,3584,0.029727111260096233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,32,3584,0.009131555755933126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,51200,0.24892799059549967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,3072,0.0736008882522583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,32,0.20492445098029244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,128,0.0028533333291610083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,2048,0.0064204442832205035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,8192,0.01797777745458815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,32,0.019525332583321463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,768,0.02104266650146908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,3072,0.012598222328556908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,8192,0.022543110781245764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,51200,0.3540959888034397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,5120,0.017449777987268236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,5120,3584,0.01167022188504537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,16384,12288,0.5838106473286947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,32,0.0026480000880029467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,5120,0.029888888200124104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,16384,0.052347554100884326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,3584,0.2476888762580024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,8192,0.043041778935326465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,3072,0.0211173329088423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,12288,0.03033066789309184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,12288,4.351778666178386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,128,0.008048888709810045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,1536,0.3411120043860541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,7168,0.23398666911655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,32,0.031311998764673867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,64,65536,0.10630666547351414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,8192,0.012793777717484368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,5120,0.39962487750583225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,1536,0.05975466966629028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3072,3584,0.12397421730889215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,7168,0.06543199883566962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,8192,0.15967733330196804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,1024,0.9444559945000542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,16384,0.060248891512552895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,3072,0.018013333280881245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,51200,0.6218693521287706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,12288,33.75261773003472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,10240,0.08018044630686443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,4096,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,256,0.0032951111594835916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,4096,7168,0.023560000790490046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,16384,0.7772666613260905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,8192,0.04534222351180183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,3584,0.007701333198282454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,3072,0.07297333081563313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,6144,0.6157760090298122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,16384,0.007637333538797166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,1024,0.056030220455593534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,65536,0.08831199672487046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,768,0.42449333932664657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,128,0.005333333379692501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,8192,0.026425777210129633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,512,256,0.007756444315115611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,1536,0.034880889786614314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,16384,0.08748355838987563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,5120,0.07816533247629802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,6144,0.2432017856174045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,2560,0.014871999621391296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,64,0.06035466988881429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1536,6144,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,32,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,8192,0.2563689019944933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,3072,0.057586669921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,51200,0.2962248855166965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,4096,0.07724444733725654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,4096,0.0477422210905287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,6144,0.14844444062974718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,8192,1.8039928012424047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,1024,0.010598222414652506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,128,0.003830222205983268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,512,0.01651733285850949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3584,2560,0.07068533367580838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,768,0.00976533359951443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,10240,0.032624890406926475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,128,0.002749333365095986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,8192,0.03681244452794393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,10240,0.09393422471152411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,4096,1.4226373036702473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,12288,0.3291253248850505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,8192,0.0606844425201416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,10240,1.568953726026747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,12288,0.01980622278319465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,51200,0.10031466351615058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,5120,0.021583999196688335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,8192,0.05669066641065809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,7168,0.022449778185950384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,10240,0.016342222690582275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,32,0.04389333393838671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,65536,0.3650791115230984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,6144,0.022109333011839125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,256,0.007448888487286038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,2048,0.022781333989567224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,10240,0.9834453794691297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,32,0.02889866630236308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2048,65536,0.1309537755118476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,1024,0.005767111149099138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,10240,0.006684444430801604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,2560,0.48740355173746747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,8192,0.37646132045321995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,4096,0.01917777789963616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,256,0.01091644416252772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,1024,0.014205333259370593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,5120,0.02587911155488756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,7168,0.03543022274971008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,256,0.0047155556579430895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,4096,0.0420595539940728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,12288,0.4116862350040012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,128,0.005306666509972678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,512,0.006131555471155379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,64,0.003061333257291052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,2048,0.10094577736324734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,768,0.004269333349333869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,512,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,8192,0.011875555747085147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,51200,0.7206577724880643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,65536,0.07641688982645671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,7168,0.013531555732091268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,10240,0.08873866664038764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,65536,0.1291546689139472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,512,0.0181048893266254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,51200,0.8009733094109429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,2048,1.1921858257717557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,5120,0.020089776979552377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,10240,0.06686755683686998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,65536,1.509414248996311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,2048,0.012190221912331052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,4096,0.010831111007266574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,3584,0.022783999641736347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,1536,0.017675555414623685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,1024,0.029532445801628962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,256,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,64,0.013548444542619916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,2048,0.012457778056462606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,4096,0.012098666694429187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,1536,0.007271111011505127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,32,0.007629333270920648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,8192,0.042486220598220825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,2560,0.02088800072669983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,32,0.006852444261312485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,1536,0.02340000040001339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,2560,0.1699297825495402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,512,0.00638133328821924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,2560,0.015059555570284525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,512,0.005961777849329843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,4096,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,64,0.025096888343493145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,768,0.02956977817747328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,128,0.00299822228650252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,8192,0.13399644692738852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,2048,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,10240,0.011780444118711682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,4096,0.010814222196737925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,3072,0.018391110830836825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,128,0.012080888781282636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,2048,0.011347555451922946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,8192,0.014783110883500842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,32,0.03492533498340183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,2048,0.12158221668667263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,512,0.0069013333155049225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,51200,0.08709777726067437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,32,0.002554666664865282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,65536,0.22272356351216635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,12288,0.08427200052473281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,1536,0.017419555120997958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,2560,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,16384,0.02197955548763275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,7168,0.10823200146357219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,1024,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,51200,0.055107557111316256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,12288,2.1884356604682074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,8192,0.031049778064092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2560,51200,0.09679733382331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,12288,1.0802497863769531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,7168,32,0.0032666667054096856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,2560,0.016064888901180692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,2048,0.03370755579736497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,32,0.01902222302224901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,8192,0.02516355613867442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,2048,0.22139733367496064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,1024,0.016214221715927124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,8192,0.05333955420388115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,512,0.04288444585270352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,3584,0.06249599986606174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,8192,0.011734222372372946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,65536,1.1232577429877388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,1536,0.1598151127497355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,65536,0.08456977870729233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,2048,3.805447048611111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,16384,2.6549769507514105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,3072,0.05677066908942329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,32,1024,0.004757333132955763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,10240,0.017609778377744887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,32,0.003432000055909157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,1536,0.007623111208279927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,5120,0.0843813353114658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,2560,0.038522667355007596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,256,0.005106666849719154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,512,0.01757244434621599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,32,768,0.007977777885066139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,2048,0.007239999870459239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,16384,0.29604800542195636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,4096,0.0607137746281094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,1024,0.006688000013430913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,2048,0.008375110725561777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,1536,0.005692444327804778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3072,2048,0.009869332942697737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,16384,0.04366133279270596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,1536,0.05197688937187195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,1536,0.86917511622111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,3584,0.029896888468000624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,1536,0.008482666479216682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,32,0.006365333166387346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,128,0.04311022162437439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,65536,0.25513243675231934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,51200,0.33068709903293186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,768,0.007752888732486301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,3584,0.06661688619189791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,1024,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,5120,0.03942666782273187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,1024,0.003961777935425441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,768,128,0.0042204442951414315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,10240,0.02019377715057797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3072,10240,0.12279378043280707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,12288,0.017271111408869427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,51200,2.153336842854818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,128,0.004987555659479565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,3584,0.15610399511125353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,10240,0.1700079970889621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,5120,1024,0.007849777738253275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,4096,0.0302782224284278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,32,0.0059368887709246735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,32,0.012440000143316058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,5120,0.24755377239651152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,3584,0.15124800470140246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,6144,0.01940711173746321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,16384,0.12389955255720352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,1024,0.12949511739942762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,768,0.006217777729034424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,768,0.007781333393520779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,64,0.0026133333643277488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,768,0.06817689206865099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,1024,0.005852444304360284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,2560,0.007700444095664554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,6144,0.0705617798699273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,256,0.003559999995761447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,1536,0.014783110883500842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,6144,0.044450667169359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,512,0.18010577890608045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,256,0.0052871112194326185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,512,0.0033404444240861465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,2048,0.01850666602452596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,8192,0.21480177508460152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,12288,0.006421333385838403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,64,0.01163200040658315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,6144,0.30234845479329425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,6144,16384,0.04534844557444254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1536,512,0.005072000126043956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,1536,0.058872891796959766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,2560,0.01513777838812934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,64,0.0024666666156715816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,5120,0.02623822291692098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,16384,0.03358666764365302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,3072,0.04101422097947862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,2560,0.35369155142042374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,65536,0.0696915586789449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,768,0.022608889473809138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,6144,0.16227822833591038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,7168,0.01422755585776435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,2048,0.012574222352769641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,64,51200,0.2179137733247545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,512,0.0031999999450312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,8192,0.025715554753939312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,512,0.005484444399674733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,1536,0.007096889118353526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,16384,0.1629342238108317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,512,0.01370044466522005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,4096,0.021964444054497614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,12288,0.14811110496520996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,64,0.004266666869322459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,1024,0.006164444403515921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,6144,0.02349866595533159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,51200,0.44607289632161456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,3072,0.021486222743988037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,1536,0.013050666285885705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,2048,0.039543999565972224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,1024,0.00369422220521503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,32,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,65536,1.8436720106336806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,1024,0.006805333412355847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,12288,0.8166640069749621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,2560,0.024248000648286607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,256,0.0047662220895290375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2560,16384,0.26179644796583385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,65536,0.028816888729731243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,256,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,1024,0.03547822104560004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,256,0.003519111209445529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,64,0.006064888917737537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,5120,128,0.005647999958859549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,2560,0.034522665871514216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,4096,0.014456889695591398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,16384,0.025998221503363714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,4096,0.006043555421961679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,6144,0.11042310794194539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,51200,0.022750222020679053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,16384,4.152805328369141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,10240,0.05072355601522657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,12288,0.01798488861984677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,8192,0.45203553305731875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,65536,7.413893805609809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,8192,0.04695644312434726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,2048,0.007877333296669854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,128,0.009303111169073317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,7168,0.2531590991550022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,16384,0.026901332868470088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,64,0.00740799970097012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,8192,0.10058399703767563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,128,0.0058328890138202244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,128,0.007998222278224098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,7168,1.3490240308973525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,7168,0.2966017723083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,5120,0.032644444041781955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,6144,0.5292035738627116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,5120,0.01315822203954061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3584,256,0.02127911150455475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,1024,0.007755555212497711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,1536,0.03871733281347487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,2560,0.00927111092540953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,4096,0.020449777444203694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,6144,0.013460444079505073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,256,0.005526222288608551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,16384,1.297027587890625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,51200,0.24704355663723418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,512,0.005971555494599872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,10240,0.33984888924492734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,4096,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,1024,0.009630222287442949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,2560,0.023688889212078516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,12288,0.04234666625658671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,6144,0.020826665891541373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,1024,0.03417955504523383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,32,0.006592888798978593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,3584,0.009443555441167619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,8192,0.01571999986966451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,5120,0.04815377791722616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,7168,0.2898426585727268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,7168,0.02696000039577484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,6144,0.016158221496476066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,32,0.0032915555768542816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,128,4096,0.018093332648277283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,6144,0.03464444478352865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,256,0.005703111075692707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3584,8192,0.20045067204369438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,8192,0.02077866593996684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,16384,1024,0.11400177743699814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,65536,0.48233334223429364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,8192,0.23422222667270234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,7168,0.12998399469587538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,7168,0.011483555866612328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,12288,0.04474488894144694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,768,0.00720888872941335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,7168,0.1927920050091214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,128,0.024166221419970196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,16384,0.022071111533376906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,256,0.019913777709007263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,64,7168,0.007394666473070781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,1024,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,64,0.08132533232371013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,32,0.017874666386180453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,7168,0.01698844465944502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,3584,0.11132355531056722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,1024,0.0826079977883233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,12288,0.03550666570663452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,1536,0.014520888527234396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,8192,7168,0.043056888712777026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,5120,0.015341333217091031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,12288,0.024021334118313257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,12288,0.04808977908558316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,1536,0.004690666579537922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,12288,0.15064356062147352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,3072,0.007649777664078607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,8192,0.006890666567616993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,128,0.004420444369316101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,4096,0.012270222107569376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,5120,0.009949333137936061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,12288,0.10287377569410537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,1536,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,16384,0.0928728911611769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,16384,0.031045334206687078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,16384,3.3315974341498484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,64,0.011564444336626263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,65536,5.746023813883464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,1024,0.032356444332334734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,2560,0.00814222213294771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,10240,0.008975110948085785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,1024,0.005344888816277186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,12288,0.0069013333155049225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,2560,0.010100444157918295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,2560,0.006646222124497096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,5120,6144,0.21583288245730928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,256,0.005604444278611078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,1024,0.3229431046379937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,2560,1.7271315256754558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,16384,0.353205336464776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,2560,0.006504000061088138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,4096,0.07582933372921415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,3072,0.014510222607188754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,3072,0.3839075565338135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,512,0.011891555454995898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,5120,0.008442666795518663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,1024,0.008782222039169734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,1536,0.013916444447305469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,65536,64,0.026018665896521673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,1024,0.03635822070969476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,32,0.0070879997478591064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,7168,0.42869777149624294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,10240,0.2727466689215766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,32,0.00304888892504904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,512,0.010567111273606619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,10240,0.04131733377774557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,2048,0.007073777417341868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,5120,0.011681777735551199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,65536,0.06541777981652154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,768,0.004239999999602635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,4096,3072,0.011377777490350934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,10240,0.024059555596775476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,64,0.00835111074977451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,65536,0.2580213281843397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,1024,0.007139555282062954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,32,0.04263822237650553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,2048,0.005602666487296422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,256,0.0033866665843460294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,4096,0.007879999776681265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,12288,0.041249778535630964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,1536,0.005943111247486538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,51200,128,0.013006221916940479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,4096,0.6191128624810113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,768,0.0870506697230869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,1024,0.048307554589377515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,1536,0.18268356058332655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,8192,0.13507733080122206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,64,0.006175111151403851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,5120,0.02889866630236308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,1024,0.0834088921546936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,51200,0.9747769037882487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,32,5120,0.0063973334100511335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,768,0.009948444035318162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,16384,0.040752000278896756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,2560,0.00831200016869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,768,0.04768888817893135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,256,0.26185867521497935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,768,0.00462311092350218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,6144,0.35393155945671934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,16384,0.15158489015367296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,512,0.04885688755247328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,768,0.051111999485227794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,5120,0.023893333143658105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,5120,0.03893866803910997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,1024,0.07068000237147014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,7168,0.3558071189456516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,128,0.0033724444607893624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,32,0.006040888941950268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,256,0.0105608892109659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,8192,0.011433777709801992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,3072,0.01666933298110962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,512,0.009159999589125315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,6144,0.013414222333166333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,4096,0.011718221836619906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,2560,0.034429334931903414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,3584,0.008707555631796518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,32,0.006887110984987683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,8192,0.025891555680169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,16384,0.041144887606302895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,4096,65536,0.13625778092278376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,6144,0.057929780748155385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,5120,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,64,0.005416000054942236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,2560,0.017722666263580322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,4096,0.007986666427718269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1024,2048,0.005484444399674733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,51200,2.4771866268581815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,10240,0.24534577793545195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,5120,0.3208693398369683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,6144,0.012401777837011548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,7168,0.014242667290899487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1024,1536,0.021327111456129286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,4096,0.05399822195370992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,3584,0.14577421877119276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,256,0.02216888964176178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,12288,0.09364089038636948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,512,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,5120,0.01833777791923947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,6144,0.006224000205596288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,256,0.004746666798988978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,8192,0.013908444179428948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,51200,0.08902666966120402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,2560,0.026521777113278706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,7168,0.017789334058761597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,128,0.004632888982693355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,1536,0.006480888773997624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,12288,0.5161173078748915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,3072,0.012860444684823355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,5120,0.027702222267786663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,6144,0.017468444175190397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,8192,0.022608889473809138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,768,0.048956443866093956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,3072,1.990738762749566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,16384,0.7796719868977865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,512,0.05288088983959622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,12288,0.167303999265035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,1536,0.011564444336626263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,2048,0.023565333750512864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,1024,0.012605333493815528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,1024,0.004024888906213972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,4096,0.012178666889667511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,51200,0.08607555760277642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,512,0.004911111046870549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,512,0.004476444588767158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,3584,0.030978666411505804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,3072,0.10545777612262303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,5120,0.05968266725540161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,12288,0.05806755357318454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,51200,18.76708984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,10240,51.614963107638886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3584,256,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,5120,0.4517342249552409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,65536,0.08557422293557061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,16384,0.2255484528011746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,10240,3.516272862752279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,768,12288,0.02274844381544325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,5120,0.3343511157565647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,5120,0.01886577738655938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,128,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,256,0.015124445160230001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,768,0.017286222842004564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,16384,0.10230755805969238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,2560,0.07267466518614027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,10240,0.09632088740666707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,128,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,1024,0.006299555715587404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,256,0.012501333322789935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,768,0.0035991109907627106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,3584,0.7912649048699273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,51200,0.07261511352327135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,65536,0.11516711446974014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,4096,0.008360889222886827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,1024,0.003976888954639435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,5120,0.00982844415638182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,768,0.08338577879799737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,1024,0.009782222410043081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,2560,0.00776444458299213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,512,0.007988444632954067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,1536,0.006919999917348226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,512,0.00907466643386417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,32,0.022138666775491502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,10240,0.030224889516830444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,16384,0.6905493206448026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,16384,0.018910222583346896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,10240,0.07992888821495904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,3072,0.012278222375445895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,10240,0.26563554339938694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,32,0.05753778086768257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,8192,0.5978817939758301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,512,0.007480888730949826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,7168,0.08580355511771308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,65536,12288,0.22259200943840873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,7168,0.009385777844323052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,256,3584,0.006747555401590135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,6144,0.02321333355373806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,128,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,256,0.01110666659143236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,32,0.024043555061022442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,12288,0.014273777604103088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,10240,0.706615130106608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,768,0.012849777936935425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,4096,32,0.019352000620630052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,768,0.03372622198528714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,1024,0.0068853331936730286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,2048,0.007505777809354994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,6144,0.5712178018358018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,51200,2.6566248999701605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,51200,0.0891759991645813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,256,0.005425777700212266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,1024,0.004571555389298333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,128,0.0030542222989930045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,3584,4.398990207248263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,3072,0.22081955273946127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,4096,0.006308444258239534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,3072,0.010317333042621613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,1536,0.0063786668082078295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,3584,0.06342489189571805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,12288,0.059911111990610756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,256,0.003067555526892344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,128,0.00518133325709237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,51200,65536,0.9084506564670138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,7168,0.017815111411942374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,6144,0.049954666031731494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,3072,0.017280000779363845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,4096,0.013632000320487551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,1024,0.006167111297448476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,64,7168,0.01109688894616233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,51200,0.17834222316741943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,3584,0.013056000073750814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,1536,0.01315822203954061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,3072,0.037294222248925105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,7168,0.21470310952928331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2048,128,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,8192,0.011869333684444427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,512,8192,0.2223351001739502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,768,0.13649955060746935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,128,0.0027297778675953546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,2560,0.07267466518614027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,512,0.027718222803539697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,128,0.020478222105238173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,128,0.01314133322901196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,256,0.033709334002600774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,2560,0.007357333269384172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,4096,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,4096,0.016135111451148987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,256,8192,0.007867555651399825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,8192,0.011751111182901593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,32,0.03383377856678433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,768,0.010507555471526252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,512,0.0034799998005231223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,512,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,3584,0.013749333719412485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,768,0.04097155398792691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,2048,0.007575111256705389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,32,0.029917332861158583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,10240,0.03981511129273309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,128,2560,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,51200,0.012300444145997366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,32,0.004863111095296012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,6144,0.9498631159464518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,3584,0.015899555550681222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,1536,0.04641600118743049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,12288,0.07570577992333306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,2048,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,512,0.014651555154058667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,2048,0.012607999973826937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,32,0.0026337777574857077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,5120,0.010835555692513784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,64,0.02308977809217241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,5120,0.010244444840484196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,5120,0.0310515562693278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,16384,0.031439112292395696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,16384,0.05467999974886576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,7168,0.01610488858487871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,1536,0.009210666848553551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,10240,0.015799111790127225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,2560,0.01835644410716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,12288,0.20388266775343153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,3072,0.045349333021375865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,2560,0.012011555333932241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,6144,0.009864889085292816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,768,0.10005866818957859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,51200,0.0369102226363288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,512,0.005492444253630108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,2560,0.007118222614129384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3584,256,0.00462666650613149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,5120,3072,0.012523555921183692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,6144,0.01717155509524875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,3584,0.012588444683286877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,3584,0.009215110705958473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,768,0.006099555641412735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,256,0.00601244428091579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,256,128,0.004953777624501122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,12288,0.015601777368121676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,3584,0.02072888943884108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,3072,0.007793777518802219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,4096,0.2532622284359402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,51200,0.0557840002907647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,6144,0.014676445060306124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,4096,1.5845830705430772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,3584,0.017442666822009616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,1024,0.005886222339338726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,12288,0.43207822905646426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,8192,0.1755759980943468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,256,0.003690666622585721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,256,0.0047893333766195504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3584,2048,0.08564444382985432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,16384,0.12100444899664985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,64,0.002557333351837264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,2048,0.016890666551060148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,8192,0.03242044316397773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,4096,0.008316444853941599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,2048,0.01202755586968528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,768,0.017886221408843994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,512,0.00554577757914861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,128,0.0063324446479479475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,6144,2.654378679063585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,16384,0.0955431130197313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,768,0.031983110639784075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,768,0.01901955571439531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,4096,0.013487999637921652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,2048,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,8192,0.10537511110305786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,16384,0.025441777375009324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,8192,0.008054222497675154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,10240,2.172337849934896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,128,0.04340711236000061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,5120,0.02776977750990126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,5120,0.009582222335868413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,4096,0.20963822470770943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,2560,0.09713421927558051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,256,0.004100444416205089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,64,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,256,0.057723555299970836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3072,768,0.01371022231049008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,4096,0.020039111375808716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,4096,1.7921138339572484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,3072,0.052453332477145724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,1536,0.04432533184687296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,2048,0.00927111092540953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,65536,0.11283822854359944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,1536,0.03553955422507392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,51200,0.21001956197950575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,8192,0.08141777912775676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,8192,0.030990223089853924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,1536,0.1196391052669949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,7168,0.0528675549560123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,32,12288,0.027183110515276592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,512,0.007035555938879649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,1536,0.02414488957987891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,12288,0.28426843219333225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,10240,0.19614222314622667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,64,3072,0.008783111141787635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,4096,0.03422666589419047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,5120,0.17053243849012586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,768,0.0034640000926123727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,512,0.0034088889757792154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,16384,0.09158844417995876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,256,0.007491555478837755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,7168,0.03441866570048862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,4096,0.039928889936871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,1536,0.03172533379660712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,16384,0.4452551205952962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,8192,0.04258666767014397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,6144,0.21606132719251844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,512,2048,0.017113778326246474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,5120,512,0.020609777834680345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,1024,0.005585777676767773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,128,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,3072,0.08250666989220513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,2048,0.013117333253224691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,12288,0.24433512157864043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,8192,0.015013333823945789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,16384,0.1515315506193373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,10240,0.07133866680992974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,10240,0.024970667229758367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,8192,0.0074755557709270054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,16384,0.0306000014146169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,12288,0.04649066593911913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,8192,0.32285067770216197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,12288,0.006931555353932911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,2048,0.06778666708204481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,768,0.006269333263238271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,1536,0.018623110320832994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,16384,0.1729208893246121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,3584,0.24506399366590711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,6144,2.8741465674506297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,64,0.0036186666952239144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,1536,0.15460266007317439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,10240,1.801161872016059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,3072,0.02884888980123732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,2048,0.01651822196112739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,51200,34.02244398328993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,2560,0.26599645614624023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,5120,0.010934222075674268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,3072,0.24530577659606934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,3072,0.006727111008432176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,128,0.011128889189826118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,1536,0.006135111053784688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,256,12288,0.3033466604020861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3072,4096,0.11763111750284831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,10240,0.04477599925465054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,8192,0.07251022259394328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,6144,0.01808533403608534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,12288,0.020040889581044514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,64,0.006004444426960415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,7168,0.04889600144492256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,51200,2048,0.03649955656793382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,3584,0.42903465694851345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,1536,0.006862222320503659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,1024,0.005387555393907759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,16384,0.030557334423065186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,256,0.0358533329433865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,8192,5.850653330485026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,128,0.005461333526505365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,512,0.01389955480893453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,32,0.010595555934641095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,7168,0.10338311062918769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,5120,0.012179555164443122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,768,0.005470222069157495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,4096,0.0320355561044481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,768,0.0058666666348775225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,1024,0.03161333335770501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,5120,0.01404355631934272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,3584,0.2178515593210856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,3072,0.007146666447321574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,10240,0.04516622092988756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,64,0.005740444279379315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,16384,0.03961955507596334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,3072,0.010798222488827176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,8192,0.8149929046630859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,1024,0.006535999890830781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,64,0.0034142221427626083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,51200,1.1532310909695096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,4096,0.01255644443962309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,32,0.005027555757098728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,128,1536,0.004490666505363253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,3072,0.007418666448858049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,3072,0.009114666945404476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,7168,0.33664711316426593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,51200,0.07916355795330472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,8192,0.18059378200107148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,7168,1024,0.11205244064331055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,512,0.026387555731667414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,128,0.011437333292431302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,64,0.00315022220214208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,768,0.0064284445510970215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,2560,0.09585244125790066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,2560,0.023795555035273235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,12288,0.01739911072784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,10240,0.024884444144037034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,32,0.0026977778308921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,8192,0.030549334155188665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,64,0.008197333249780867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,128,16384,0.023457777169015672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,128,0.006153777655627992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,2048,0.03482133481237624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,1536,0.009674666656388177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,3072,0.006699555450015598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,2048,0.010117332968446944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,2560,0.03261333372857835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,4096,0.02346222268210517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,2560,0.014395554860432943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,64,0.007450666692521837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,6144,0.011215111447705163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,4096,0.007912889122962952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3072,128,0.004283555679851108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,51200,2.4579137166341147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,10240,0.3030915525224474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,1024,0.007887110941939885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,12288,0.028939555088678997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,256,0.0046880000995265115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,128,0.008280889027648503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,16384,0.03240711159176297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,64,256,0.0031155554784668815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,1024,0.03567022085189819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,16384,0.10035644637213813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,2560,0.6648649109734429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,3072,0.2917582194010417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,16384,0.020966221888860066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,3584,0.281912883122762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,65536,0.07125066386328803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,64,0.0064275554484791225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,2048,0.18561600314246285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,1024,0.010552888943089379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,7168,0.12447910838656956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,6144,0.010191111101044549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,12288,0.042240000433391996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,3072,0.15227199925316706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,128,0.0031884445084465873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,128,0.007191999918884701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,12288,0.02751644452412923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,5120,0.016275554895401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,10240,0.3260435528225369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,2048,0.009199110998047723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,51200,128,0.009390222529570261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,5120,0.08960177501042683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,3584,0.010568888651000129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,8192,0.033928887711630926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,1024,0.00498844434817632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,768,0.013005333642164866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,128,0.015249777171346875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,5120,0.015422221687104968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,5120,0.041660444604025945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,51200,0.14517688751220703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,65536,0.10156355301539104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,8192,0.020833777056799993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,51200,0.09806044234169854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,16384,0.03279111120435927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,2560,0.009461333354314169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,10240,0.040256887674331665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,7168,0.04184000028504265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,5120,0.36574755774603945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,512,0.008525333470768398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,512,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,7168,0.032256000571780734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,2048,0.047162665261162647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,6144,0.01956977778010898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,16384,0.061611553033192955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,7168,0.20592978265550402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,1024,0.025289777252409194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,32,0.030477331744299993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,3584,0.012674666941165924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,128,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,7168,0.22536532084147134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,1024,0.012906666431162091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,512,0.006158222340875202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,10240,0.02123822271823883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,51200,0.07618666357464261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,16384,0.22232532501220703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,65536,0.0969724456469218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,256,0.010695111420419482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,5120,0.0240657776594162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,12288,0.01789066692193349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,256,0.002986666642957263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,512,0.02342844506104787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,3584,0.008807111117574904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,768,0.013704000247849358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,6144,0.07043733199437459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,65536,0.043474667602115206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,5120,0.009101333717505137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,3584,0.04112711217668321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,6144,0.03545688920550876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,512,512,0.008295111358165741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,51200,0.755591975318061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,2048,0.04588977826966179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,3072,0.006471111128727595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,65536,16.905643039279514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,51200,8192,4.214829338921441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,5120,0.06704533100128174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,3584,0.025096888343493145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,7168,0.19522044393751356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,5120,0.3477671146392822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,12288,0.030718220604790583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,10240,0.037436445554097496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2048,6144,0.07441422012117174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2560,4096,0.8610000080532498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,51200,0.10931466685401069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,3072,0.011457777685589261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,2048,0.03789244426621331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,6144,0.018613333503405254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,1024,0.044567998912599355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,768,0.006170666466156642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,256,7168,0.006383111079533895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,12288,0.8988204532199435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,512,0.003942222230964237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,7168,0.0539991127120124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,2048,0.08955022361543442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,4096,0.01590666671593984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,2560,0.007836444510353936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,1024,0.004808889081080754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,128,0.05023911264207628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,64,0.00434666665063964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,128,0.004872000051869286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,64,0.0027377777215507296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,12288,0.13821510473887125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,16384,0.02145066691769494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,3072,0.03304800060060289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,5120,0.09816088941362168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,16384,0.07817600170771281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,51200,0.3392008940378825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,65536,0.08377689123153687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,10240,0.025624889466497634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2560,65536,0.4158951176537408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,1536,0.03402933478355408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,128,0.005057777795526716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,5120,0.06627288791868421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,768,0.03858933183881972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,16384,0.12498311201731364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,3072,0.10447111394670273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,7168,0.06937066713968913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,65536,0.015326221783955893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,256,0.006478222293986215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2560,51200,1.1500088373819988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,3584,0.3003608915540907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,6144,0.13993244700961643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,3584,0.0103511114915212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,7168,0.33624890115525985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,7168,0.1647635565863715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,64,0.010238221950001186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,12288,0.08482666810353597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,6144,1024,0.006372444331645966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,32,0.003068444422549672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,3584,0.027058665951093037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,16384,13.514645046657987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,10240,0.33196621470981175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,8192,64,0.1030728883213467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,512,0.005397333453098933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,512,0.026159110996458266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,7168,0.02842666705449422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,64,0.0052328887912962176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,16384,0.1825102170308431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,12288,0.015594666202863058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,512,0.006380444599522485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,7168,0.06269333097669813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,7168,0.01959288948112064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,256,0.02882044514020284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,3072,0.08833689159817165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,6144,0.0262097782558865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,2048,0.005809777726729711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,6144,0.2547875510321723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,8192,0.14378400643666586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,256,0.004359999878538979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,128,0.0028391112056043413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,6144,0.045369777414533824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,5120,0.037323554356892906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,512,0.018771555688646104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,5120,0.014984889162911309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,5120,0.04755111204253303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,256,0.0031724443866146933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,12288,0.2311075528462728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,51200,0.23175644874572754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,64,0.010942222343550788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,512,0.6000257598029243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,51200,0.06084799766540527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,2560,0.028435554769304063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,16384,0.006684444430801604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,512,0.0034799998005231223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3584,1024,0.004911111046870549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,256,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,4096,16384,0.04545066754023234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,32,0.014142221874660917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,1024,0.13660621643066406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,10240,0.04438933398988512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,4096,0.008359111017651027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,768,0.006129777679840724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,768,0.010339555641015371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,12288,256,0.006355555521117316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,16384,1.7340870963202581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,128,0.03478044602606032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,10240,0.1965013345082601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,5120,0.017180444465743173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,64,0.007349333001507654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,8192,0.3345404465993245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,1536,0.013536888692114087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,51200,0.05288266804483202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,12288,0.027080888549486797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,12288,0.20312089390224883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,6144,0.06234221988254123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,51200,0.21531911691029867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,4096,0.5542337629530165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,12288,0.14087911446889242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,2048,0.1631635559929742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,7168,0.14188710848490396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,10240,0.17221689224243164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,256,0.014680888917711047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,128,0.014283556077215405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,32,0.00535022219022115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,5120,0.18367555406358507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,32,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,3584,0.007900444169839224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,6144,0.2957831223805745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,512,0.028326223293940227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,16384,1.2541102303398979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,51200,0.05365777677959866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,3072,0.02104444470670488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,256,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,65536,0.3698586622873942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,2560,0.01475911173555586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,12288,0.015657777587572735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,256,0.047388444344202675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,2048,0.15522844261593288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,51200,0.06369155645370483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,768,3584,0.00795022232664956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,2560,0.007061333292060428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,128,0.03314222229851617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,8192,0.022999111149046157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,1536,0.04651111033227709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,6144,64,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,8192,51200,0.14533510473039415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,2048,0.0860533316930135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,3072,0.013288000391589271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,8192,4.966411590576172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,3072,0.7013306617736816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,2560,0.030447999636332195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,2560,0.2924213409423828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,6144,0.015354666444990369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,256,0.006859555426571105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,4096,1.257081773546007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,32,0.016113777955373127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,5120,0.0999422205819024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,5120,0.24885333908928764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,4096,0.012316444681750404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,65536,0.13999822404649523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,768,0.3670933246612549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,3072,0.010520888699425591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,32,0.004848000076082018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,128,0.005151111218664381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,1536,0.006180444525347815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,8192,0.014166222678290473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,512,0.0036106668412685394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,16384,0.03344088792800903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,3584,0.02212000058756934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,7168,0.43538665771484375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,3072,0.09589955541822646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,2560,0.013050666285885705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,10240,0.03262933426433139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,3072,0.006437333507670297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,4096,0.045904888047112354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,7168,0.03941599859131707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,16384,0.1313520007663303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,5120,8192,0.06103377872043186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,1024,0.007721777591440413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,256,0.1441377798716227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1024,512,0.008000888758235509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,32,0.010834666589895884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,16384,0.30734843677944607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,2048,0.005024000174469418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,4096,2048,0.06050222449832492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,51200,0.11492799388037787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,768,0.05556266837649875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,6144,0.02179644505182902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,1536,0.01628977722591824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,7168,2560,0.14744355943467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,8192,0.13358133369021946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,64,0.006989333364698622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,7168,16384,0.03972444600529141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,1024,0.0758497781223721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,8192,0.012072888513406118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,12288,0.505607975853814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,2560,0.06327999962700738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,10240,0.060283554924858936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,4096,0.021387555532985266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,1024,0.045590221881866455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,1024,0.03689688775274489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,64,0.009738666315873465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,12288,0.01575555569595761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,64,0.004933333231343163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,512,0.024870221813519795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,4096,0.016171556380059984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,4096,2.070037417941623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,6144,0.03198755449718899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,512,0.005559111220969095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,5120,0.017076444294717576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,7168,0.025928889711697895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,51200,1536,4.021131727430555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,7168,0.6019413206312392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,2048,0.015115555789735583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,4096,0.11744533644782172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,64,0.03144711256027222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,64,0.0026488889836602737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,65536,2.151664945814345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,5120,0.03970311085383097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,32,0.002620444529586368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,65536,0.029937777254316542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,2560,0.011131555669837527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,1024,0.004451555510361989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,4096,0.01479822231663598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,4096,0.02128977742460039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,1024,2.0294257269965277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,16384,0.04801688922776116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,16384,0.81805419921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,1024,0.006507555643717448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,3584,0.027448889282014634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,512,0.02879022227393256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,1024,0.018925334016482036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,4096,0.030119998587502375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,16384,0.09729866849051581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,3072,0.038784000608656145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,2048,0.8628124660915799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,7168,0.023050665855407715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,32,0.002579555536309878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,1024,0.02825333343611823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,512,0.015947555502255756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,1024,0.009181333084901174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,7168,0.020308444897333782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,12288,0.7004959848192004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,768,0.05213510990142822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,8192,1.3274737464057076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,1024,0.007744000189834171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,2560,0.008598222500748104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,6144,0.12045155631171332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,1536,0.008639111287064022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,2560,0.15647378232744005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,2048,0.27370309829711914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,7168,4.653415256076388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,768,128,0.004400889078776042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,1024,0.08343911170959473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,256,0.004402666870090696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,12288,0.22066577275594076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,12288,0.0471955570909712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,10240,0.30049777030944824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,256,0.0029208889851967492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,3584,0.01111288865407308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,256,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,12288,1.6965946621365016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,5120,1.778289794921875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,128,1024,0.005629333357016246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,8192,0.01791200041770935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,3584,0.008593777815500895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,8192,0.32353332307603627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,65536,0.4667546484205458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,1536,0.005912000106440649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,4096,0.009685333404276106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,4096,0.03506844573550754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,512,0.01128000020980835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,7168,0.10841066969765557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,256,0.005972444597217772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,128,0.003272888975010978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,5120,0.022314666046036616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,10240,16384,0.060137775209214955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,16384,0.0650728874736362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,16384,0.02264533274703556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2560,4096,0.02552266584502326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,2560,0.010682666467295753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,65536,0.9900293350219727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1024,65536,0.3312471177842882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,1024,0.022828444838523865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,10240,0.05717511309517754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,16384,0.013844444519943662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,65536,0.875855975680881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,7168,0.9219421810574002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,128,0.015305777390797934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,2048,0.03363644414477878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,3072,0.05323733223809136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,8192,0.008755555583371056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,6144,0.01921955578856998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,3584,0.07371999820073445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,256,0.0046817776229646476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,51200,4.286823272705078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,768,0.011607111328177981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,64,0.009098666409651438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,8192,7168,0.4393760098351373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,4096,0.009295110901196798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,256,0.0032666667054096856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,1024,0.008767999708652496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,128,0.003960888832807541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,65536,0.42753156026204425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,10240,0.32761600282457143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,1536,0.00885511106914944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,2560,0.015084443820847405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,128,0.014878221684032016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,1024,0.006161777923504512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,512,0.011336888704035016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,64,0.0058453331391016645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,2560,0.11729155646430121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,2560,0.058218664593166776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,12288,0.02619022296534644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,256,0.006592888798978593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,1024,0.007096000015735626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,12288,0.4095493422614203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,1536,0.03250755535231696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,12288,0.39663465817769367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,10240,0.06474844614664714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,128,0.002756444530354606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,3072,0.018726223044925265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,512,0.013813333378897773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,51200,0.08257955312728882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,5120,0.05778666337331136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,2560,0.014546665880415173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,12288,0.02977333466211955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,3072,0.017532444662517972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,3072,51200,1.9308213127983942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,4096,0.11573778258429633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,51200,0.8930177688598633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,5120,0.1000702248679267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,10240,0.39897600809733075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,512,0.004613333278232151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,512,0.0058444444504049085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,2048,0.006206222292449739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,51200,0.012796444197495779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,6144,0.4361591074201796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,256,0.005151999907361137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,1024,0.6553911103142632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,10240,0.009308444129096137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,65536,0.2933146688673231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,256,0.017793777916166518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,2560,0.009248889154858058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,4096,0.0955324437883165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,1024,0.007231999602582719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,1536,0.00884622252649731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,1536,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2560,6144,0.0481279989083608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,8192,0.3343653413984511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,2048,0.007171555525726742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,256,0.10139733552932739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,7168,0.10074222087860107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,64,0.00313688897424274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,3072,0.29172976811726886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,128,0.006242666807439592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,6144,8.199030558268229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,65536,5120,3.692859649658203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,2048,65536,0.8210177951388888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,12288,0.026154667139053345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,6144,256,0.003996444245179494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,7168,0.05845333470238579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,7168,0.012518222133318583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,8192,0.01715555621518029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,32,0.0031786666562159858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,6144,0.032708442873424955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,10240,0.0758320026927524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,1024,0.06808444526460436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,1024,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,32,0.09518755806816949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,768,0.0064186664919058485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,1536,0.031143999762005273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,4096,0.019321777754359774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,7168,0.016213332613309223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,1536,0.005362666729423735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,768,0.006168889088763132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,6144,0.02795377704832289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,64,0.011134222149848938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,1536,0.009164444274372524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,51200,0.22724710570441353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,2048,0.0069164443347189165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,768,0.005787555542257097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,256,0.009087111386987898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,6144,0.09068889088100857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,64,0.02775466607676612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,1024,0.006116444451941385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,6144,0.2981911235385471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,32,5120,0.0069671107663048645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,65536,0.8324942058987088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,6144,0.013035555680592855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,3584,0.02570666703912947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,16384,1536,0.018159111340840656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,1024,0.07157955567042033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1536,51200,0.0966799987687005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,12288,768,0.011215999722480774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,65536,256,0.1308177842034234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,4096,0.03730222251680162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,5120,0.39729955461290145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1024,10240,0.03189244535234239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,3072,0.015105777316623263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,7168,0.3896666632758246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,8192,0.27261511484781903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,128,0.004568888909286923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,5120,0.07486577828725179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3072,5120,0.015304888288180033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,3584,0.0809075567457411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,16384,0.27982844246758354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,7168,0.5293413268195258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,128,0.005856888989607493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,65536,13.975677490234375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,1536,0.015014222926563688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,512,0.003313777761326896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,7168,0.010945777926180096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,64,0.0031439999325407874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,12288,0.37934666209750706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,768,0.004263999975389904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,1024,0.007284444239404466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,64,256,0.0035786665976047516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,3584,0.04045955671204461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,768,2560,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,16384,0.18044800228542754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,5120,0.7539564238654243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,12288,0.04984977841377258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,3072,0.3384675449795193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,2048,0.29164621565077037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,7168,0.04621511035495334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,4096,0.015451555450757345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,2560,0.10529866483476426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,2048,0.0422435568438636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,2560,0.007685333490371704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,3584,0.5353715684678819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,51200,1.6435040367974176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,12288,0.058528886901007764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,1536,128,0.005196444276306364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,256,0.003540444291300244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,10240,1.869241714477539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,3072,0.02225333286656274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,3072,0.007862221863534715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,4096,0.06760355499055651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,2048,0.013083555632167392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,7168,0.02755911151568095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,7168,0.33262043529086643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,2048,0.09342222081290351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,65536,3.0744185977511935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,12288,0.008887999587588841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,512,0.036179555786980525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,512,8192,0.015875554747051664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,3072,0.011801777614487542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,12288,0.0638657808303833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,128,0.004949333353175057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,51200,3.1787643432617188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,3072,0.012079999678664737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,7168,0.035926222801208496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,65536,1.0658533308241103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,65536,0.07646577888064914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,512,0.016740444633695815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,2048,0.054626666837268405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,512,0.007603555917739868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,12288,0.9819066789415148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,10240,0.416152901119656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,6144,0.016703999704784818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,1536,0.05320088730918037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,4096,0.03232533401913113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,2560,0.006801777829726537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,10240,0.08366755644480388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,768,0.012530666258600024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,6144,0.02125777800877889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,128,12288,0.019813333948453266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,128,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,2560,0.006150222072998683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,768,0.018977777825461496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,6144,0.01019377758105596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,65536,0.9045066833496094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,51200,0.3146168920728895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,3584,0.012098666694429187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,16384,0.5208249092102051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,16384,10240,0.4889386494954427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,1024,0.009844444692134857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,3584,0.008430221842394935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,2048,0.007506666911972894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,16384,0.3431217670440674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,3584,1.060299555460612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,8192,0.3561066786448161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,10240,0.030229333374235366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,3584,0.010652444428867765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,4096,0.11102755864461263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,128,1024,0.005326222214433882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,65536,5.7251171535915795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,8192,0.06394044558207194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,3584,0.025587555434968736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,6144,0.012280888855457306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,5120,10240,0.187553776635064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,1024,0.004101333518822988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,128,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,256,0.013454222016864352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,512,0.006620444357395172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,5120,0.14032533433702257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,3072,0.06279644701215956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,12288,0.023099554909600153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,1024,0.20029511716630724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,3584,0.01219377749496036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,1536,0.026556443836953905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,256,0.0034088889757792154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,1024,0.008126222425036961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,8192,0.012183999849690331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,5120,0.02193244463867611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,3072,0.009518221848540837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,64,0.005787555542257097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,1024,0.008037333687146505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,1024,0.025778666138648987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,8192,0.02695911129315694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,7168,0.01053244454993142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,1536,0.004648888690604104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,4096,0.00813422186507119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,1536,0.011432888607184092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,1024,0.003965333518054751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,51200,0.13714310857984754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,51200,0.06655822197596233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,65536,2.6545066833496094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,10240,0.008344000412358178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,768,0.00942666663063897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,64,0.003079111170437601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,1536,0.3255395624372694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,64,0.03264888789918687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,32,0.0052639999323421055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,10240,0.019626667102177937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,10240,0.06052888764275444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,51200,2.616415023803711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,51200,5.001335991753472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,65536,5.28239017062717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,2560,0.009725333087974125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,4096,64,0.0032293332947625052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,10240,0.021664000219768945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,3584,0.0123831108212471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,256,0.004605333424276776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,16384,1.059931543138292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,64,0.009206222163306342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,1024,0.05661955806944105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,2560,0.013321778012646569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,64,0.003313777761326896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,5120,0.23733690049913195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,3584,0.011920889218648275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,7168,0.031121777163611516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,65536,4.24871826171875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,4096,0.010823110739390055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,256,0.013259555730554791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,51200,31.065131293402775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,7168,0.03411555621359084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,10240,0.020371556282043457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,6144,0.04450400008095635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,7168,768,0.007914666500356462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,32,0.009168888959619734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,1024,0.0447635551293691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,3072,0.156640887260437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,32,0.011190222369299995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,2560,0.011598221957683563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,51200,0.019627556204795837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,2560,0.024136889312002394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,16384,0.8051235410902234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,32,0.02624977793958452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2048,256,0.0047315553658538395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,1024,0.028063999281989202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,65536,0.09322488970226711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,128,0.0028764444092909494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,256,0.004117333226733738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,16384,0.06410133176379733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,512,8192,0.018011555075645447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,7168,0.13724799950917563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,768,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3584,3072,0.024365334047211543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,10240,0.022720888257026672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,256,0.0052622221410274506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,8192,0.03120622369978163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,7168,1.210020489162869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,256,0.03274222215016683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,51200,12288,0.1844266653060913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,51200,2.2303333282470703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,32,0.04156711035304599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,5120,0.013423110875818463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,8192,0.0355386667781406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,5120,4.472886403401692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,5120,3.7846561008029513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,8192,0.11860711044735378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,7168,0.02275377843115065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,3584,0.020706666840447318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,1024,0.01333866682317522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,16384,1.5252631505330403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,12288,5120,0.1906142234802246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,4096,0.03398133317629496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,3072,0.7506551212734647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,128,0.0031999999450312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,2048,0.15572444597880045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,12288,0.16359910700056288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,10240,0.01735466718673706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3072,32,0.007068444457319047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,12288,0.05382844474580553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,6144,0.016347555650605094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,8192,0.07353688610924615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,65536,1.8190142313639324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,10240,0.05399822195370992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,5120,0.06262311008241442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,2560,0.0561715563138326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,1536,0.010729778144094678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,256,0.005589333259397083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,512,0.011253332926167382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,768,0.2706506782107883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,64,4096,0.20355733235677084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,64,0.002955555501911375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,512,0.11219910780588786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,512,0.014075555735164218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,5120,0.01954844428433312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,7168,0.050083554453319974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,16384,0.12094933456844753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,1024,0.009886222581068674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,16384,0.08733510971069336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,3072,0.017719111508793302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,3584,0.02130488885773553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,32,0.013632000320487551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,6144,3.166462156507704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,51200,0.037584000163608126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,4096,0.2931182119581434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,128,3584,0.007310222420427535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,128,0.011787555283970304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,10240,0.026418666044871014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,3584,0.0529386666085985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,7168,0.015807999504937064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,2048,0.7696400218539767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,6144,0.015228443675571017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,3584,1.099098629421658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,256,0.008795555267069075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,3584,0.08228800031873915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,5120,0.029030223687489826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,3584,0.02548533346917894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,32,0.013153778182135688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,4096,0.020541333489947848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,1536,0.009710222482681274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,2560,0.007735110819339752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,1536,0.007311111523045435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,5120,0.021147555775112573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,2560,0.007596444752481248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,512,0.0069679998689227635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,64,0.024104000793562994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,65536,0.21129243903689912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,12288,0.3359093401167128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,32,0.02295200030008952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,1024,0.005193777796294955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,3584,1.9915396372477214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,1024,0.014860444598727755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,12288,0.4530488650004069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,7168,0.020308444897333782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,6144,0.5918133523729112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1024,5120,0.012332444389661154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,8192,0.5116071171230739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1536,256,0.005117333183685939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,32,0.008851555486520132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,768,0.03684444559945001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,16384,0.02292888859907786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,256,0.008189332981904348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,512,0.019341333044899833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,12288,0.01977244516213735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,2560,0.17917688687642416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,128,0.00674311113026407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,64,4096,0.008915555146005418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,12288,4.4669269985622835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,64,0.006464889066086874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,1536,0.021659556362364028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,7168,0.04793688986036512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,256,0.017009778155220878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,3584,0.03250577714708116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,256,0.0037644443412621817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,6144,0.6290133264329698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,65536,1.783942116631402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,3072,0.023991111252042983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,5120,0.012626666989591388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,8192,256,0.005219555563396878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,10240,0.009461333354314169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,768,0.009888889061080085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,512,0.003719999972316954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,2560,0.11111644903818767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,6144,0.2499715487162272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,32,0.048157334327697754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,1024,0.06436266501744588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,3584,0.0819466643863254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,6144,0.09503377808464898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,256,0.0046453335218959385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,10240,0.023853333459960088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,10240,256,0.005906666732496685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,12288,0.037019554111692644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,1024,0.027727999620967444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,51200,0.29255199432373047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,6144,0.03972711165746053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,65536,0.49399911032782656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,64,0.007378666765160031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,1536,0.037887109650505915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,128,0.0040942223535643685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,5120,0.1243235535091824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,12288,8192,0.054021331999037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,32,3584,0.006299555715587404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,5120,0.01738755570517646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,512,0.005906666732496685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,16384,2.64810668097602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,3584,0.013085333009560903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,6144,64,0.026479111777411565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,10240,0.017493334081437852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,256,0.0058248887459437055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,65536,0.4655289120144314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,2560,0.20398488309648302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,65536,0.07997422085867988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,768,0.0058613332609335584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,3584,0.04036622246106466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,256,0.0031537777847713898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,256,0.005351999981535806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,6144,0.0580844415558709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,51200,0.18578844600253633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,7168,0.01697422232892778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,3584,0.00925155563486947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,10240,0.006161777923504512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,6144,0.06383377975887723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,256,0.005484444399674733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,128,2048,0.006578666468461354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,3584,0.014318222800890604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2560,8192,0.014719110396173267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,10240,0.03235377868016561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,128,0.00601333338353369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,8192,0.18703822294871011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,10240,0.03176177872551812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,4096,0.01145155562294854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,6144,0.581392871008979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,65536,0.04174222217665779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,1536,0.08525511291291978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,6144,0.09359288877911037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,2048,0.018399111098713346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,512,0.05146400133768717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,65536,0.16420088873969182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,16384,768,0.010949333508809408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,512,0.004926222066084544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,5120,0.025713778204388086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,2048,0.1353200011783176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,7168,0.028384000062942505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,2560,0.008656000097592672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,7168,3.2882141537136502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,7168,0.5494026607937282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,512,0.07057155503167047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,16384,0.08882666958702935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,8192,0.8004177941216363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,1536,0.025200888514518738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,2560,0.17776177989112008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,1536,0.0074480002125104266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,1536,0.007416889071464539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,2560,0.00906133320596483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,6144,0.77328978644477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,7168,0.10515644152959187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,6144,0.06755022207895915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,3072,0.009763555394278632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,51200,0.05613244242138333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,8192,0.07575288746092054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,1536,0.011561777856614856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,8192,0.006696888970004187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,1024,0.09097688727908665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,256,0.04787466592258877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,3584,0.018089777893490262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,32,0.007818666597207388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,12288,0.030656887425316706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,51200,1.3455137676662867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,2048,0.10022222333484226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,3072,0.03516355488035414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,10240,0.20124267207251656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,65536,0.42661333084106445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,4096,0.01983022193113963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,65536,4096,13.960912068684896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,10240,65536,7.173033820258246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,10240,0.04123555620511373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,128,0.022486221459176805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,3584,0.14997155136532253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,64,0.012401777837011548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,12288,0.01035822182893753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,10240,0.5090337859259711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,768,0.3658088843027751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,16384,0.06361511018541124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,32,0.004326222257481681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,128,6144,0.01573155489232805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,2560,0.4016275670793321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,10240,1.3610364066229925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,128,0.005552000055710475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,8192,0.10202399889628093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,6144,0.5551626417371962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,5120,0.011632888681358762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,16384,2.973122702704536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,12288,65536,1.7975040011935766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,1536,0.024008888337347243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,51200,65536,10.084486219618055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2048,4096,0.016099555624855887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,256,0.010234666367371878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,10240,0.08842933177947998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,1536,0.019825778073734708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,10240,0.049754665957556836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,1024,0.010848888920413123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,128,0.0181048893266254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,2048,0.6515866915384928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,768,0.08313777711656359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,7168,0.012240000069141388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,12288,0.04325777623388502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,3072,0.010307555397351583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,12288,0.01885333326127794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,4096,0.009506666825877296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,7168,1024,0.00851288851764467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,65536,0.8832231097751193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,5120,0.1031564474105835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,3072,0.12212710910373265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,4096,0.013597333596812354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,5120,0.4778711001078288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,5120,0.019201777047581144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,64,0.0053297777970631914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,3072,0.007895111209816402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,1536,0.005964444329341252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,256,0.013984888792037964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,7168,0.3190613322787815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,2560,0.008019555773999956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,768,0.0037911112109820047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,768,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,12288,0.2660542329152425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,16384,0.28149512079026967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,768,0.008932444784376357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,1024,0.032849778731664024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,1024,0.14733866850535074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,10240,0.05784533421198527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,1536,0.019719999697473314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,6144,0.10657511154810588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,128,0.003674666500753827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,65536,0.09220355749130249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,3584,0.008524444368150499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,256,0.00810933361450831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,51200,1.2241955863104927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,7168,0.6276213328043619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,4096,0.016571554872724745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3584,10240,0.021270222134060327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,128,0.0373306671778361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,8192,0.05692178010940552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,1024,0.01015466699997584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,10240,0.015319110618697273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,8192,0.022506665852334764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,1536,0.008088000118732452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,128,0.004925333377387789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,1024,0.006543111056089401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,32,0.007803555164072249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,3072,0.011339555184046427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,128,0.02312711046801673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,6144,0.017269333203633625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,64,0.021425777011447485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,32,0.0026817777090602448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,3072,0.03182755576239692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,32,0.007567110988828871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,768,0.030851556195153132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,3072,0.012969777815871768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,1024,0.006904888898134232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,51200,2.337437311808268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,1536,0.24850932757059732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,12288,0.07036177979575263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,16384,0.019489778412712943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,768,0.0037662221325768363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,1536,0.307715548409356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,51200,0.2014604409535726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,5120,0.13004622194502088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,64,0.0030737777964936364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,2560,0.05608444743686252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,4096,0.18993955188327363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,256,0.011941333611806234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,128,6144,0.006204444501135085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,1536,0.03600177831119961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,5120,6144,0.22304089864095053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,1024,0.013081777426931592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,3584,0.014494222071435718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,256,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,10240,0.2498257690005832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,5120,0.009406222237481011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,10240,0.4862231148613824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,768,0.005469333380460739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,5120,0.09563289086023967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,7168,0.019850666324297588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,10240,0.02976977825164795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,5120,0.03652711047066583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,51200,2.6101911332872176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,7168,0.03365688853793674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,32,0.005414222263627582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,7168,0.0981760025024414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,1024,512,0.005435555759403441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,7168,0.44652356041802305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,65536,0.3215706613328722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,51200,0.27410843637254506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,512,0.005544888890451855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,32,0.006178666734033161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,3584,0.011384000380833944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,16384,0.21847732861836752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,4096,0.009855999714798397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,3072,0.10661955674489339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,3584,0.01110666659143236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,51200,0.12604888280232748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,64,0.004573333180612988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,5120,0.010024000373151567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,6144,0.015604444675975375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,2560,0.006075555665625467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,6144,16384,0.054640889167785645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,5120,0.014275555809338888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,6144,3.128200954861111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,1024,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,10240,0.13610488838619655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,6144,0.1212773323059082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,10240,0.010659555594126383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,512,0.005117333183685939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,65536,51200,39.14150661892361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,2048,0.035354667239718966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,32,0.004287999951177173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,10240,0.9775528377956815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,3072,0.032039112514919706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,1536,0.00907733374171787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,8192,0.0372613337304857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,512,0.03916088740030924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,8192,4.864877488878038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,10240,0.022592888938056097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,128,0.0034728890491856467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,1024,0.0069413334131240845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,12288,0.13238844606611463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,2048,0.054304000404145986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3584,16384,0.04702933298216926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,1024,0.006027555714050929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,8192,0.00981955561372969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3584,1536,0.00830488900343577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,51200,0.3288151158226861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,12288,0.029602666695912678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,10240,0.01624000072479248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,16384,1.2098311318291557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,8192,0.025690666503376428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,7168,0.3542337682512071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,7168,0.01549600064754486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,768,0.03190044562021891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,10240,51200,32.229590521918404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,1024,0.008757332960764566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,12288,0.2781626648373074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,12288,0.034381333324644305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1024,7168,0.07767199807696872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,3072,0.046925332811143666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,16384,0.08314488993750678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3584,1024,0.0359866652223799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,256,0.03336888882848952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,256,10240,0.11869688828786214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,16384,0.525560908847385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,7168,1.3430222405327692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,512,0.006169777777459886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,16384,0.01460088955031501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,128,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,768,0.008225777910815345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,6144,0.025602666868103877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,51200,0.2969093322753906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,768,0.0053484443989064955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,8192,0.17549688286251494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,768,16384,0.025432889660199482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,2048,0.02518044412136078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,32,0.0984613365597195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,2560,0.5084017647637261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,12288,0.0136835558546914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,5120,512,0.2040168841679891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,64,0.00259911103381051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,10240,0.018343999981880188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,1536,0.1492906676398383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,512,256,0.00665155549844106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,16384,0.0804248915778266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,2560,0.011582222249772815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,768,0.004090666770935059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,12288,0.024030221833123103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,2560,0.039352890517976545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,8192,0.43813689549763996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,16384,0.019888889458444383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,2560,0.01442311041884952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,6144,0.0768266651365492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,3584,0.06934044758478801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,12288,0.2805875672234429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,7168,3.326907475789388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,51200,12288,0.1785351170433892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,2048,0.005530666559934616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,7168,0.09008533424801296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,8192,0.46499024497138125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,16384,0.024128889044125874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,5120,0.09047289027108087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,51200,0.057890666855706104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,1536,0.010585777461528778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,64,0.009626666704813639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,32,0.008919110728634728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,2048,0.006724444528420766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,12288,0.03285155693689982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,5120,0.029839999145931665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,51200,0.21773955557081434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,8192,2.5493706597222223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2048,2048,0.009678222239017487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,256,0.004209777961174647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,512,0.011415999796655444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,2560,0.04650222261746725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,3072,0.009479111267460717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,2560,0.01722666621208191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,64,0.015250666273964776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,4096,0.03548089000913832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,2560,0.00795022232664956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,16384,0.7895155482821994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,8192,2.188507503933377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,2560,0.022453332940737408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,12288,0.07979022132025824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,10240,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,6144,0.019780443774329293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,6144,0.020590222544140287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,64,0.006304889089531369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,3072,0.07464710871378581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,6144,8192,0.02660088903374142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1024,12288,0.0780328909556071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,10240,3584,0.0169804443915685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,2048,0.043727109829584755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,4096,0.015046222342385186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,12288,0.1848693291346232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,7168,0.18266399701436362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,256,0.0036728887094391715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,5120,0.028078221612506445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,256,0.017850667238235474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,1024,0.006269333263238271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1536,3072,0.04040177663167318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1024,2048,0.013754666679435305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,7168,0.17323733700646293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,3584,0.015676443775494892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,10240,0.07249955336252849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,65536,8192,1.6948400073581273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,10240,0.22712800237867567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,16384,15.547288682725695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,4096,16384,0.37586045265197754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,8192,51200,0.9211502075195312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,16384,0.31828710767957896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,3584,0.020058666666348774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,7168,0.4455333285861545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,6144,0.06466755602094862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,6144,0.014949333336618213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,32,0.008791111409664154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,65536,2.2761190202501083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,3072,0.034953776333067156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,2560,0.020576889316240948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,5120,2560,0.04809866680039299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,12288,0.6321351263258193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,65536,3072,0.06148799922731188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,3072,0.0984577801492479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,16384,8.23004150390625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,512,0.015000889698664347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,256,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,8192,0.04083733426200019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,10240,0.10573422246509129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,768,0.007313778003056844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,1536,0.1752613385518392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,65536,0.07942222224341498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,12288,0.35642221238878036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,128,0.0027271111806233725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2048,768,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,10240,3.945312923855252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,12288,0.09377066956626044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,256,0.041476445065604314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,1024,0.0052826665341854095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,4096,0.01683555543422699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,4096,0.013487999637921652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,256,0.005342222336265776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,1024,0.02532800038655599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,768,0.14474488629235163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,8192,65536,0.23351200421651205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1536,51200,0.08572089009814793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,256,0.011472889118724398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,7168,0.03736266824934218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,2048,0.0462080008453793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,2560,0.14399022526211208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,2560,0.1493120061026679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,3072,0.10307466983795166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,3584,0.018438221679793466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,6144,0.012889777620633444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,64,0.011109333071443768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,1536,0.011358222199810876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,768,0.006120000034570694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,128,0.005202666752868229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,10240,0.056608001391092934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,768,10240,0.015496888094478183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,1536,0.5071875784132216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,1024,0.00792266676823298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,1536,0.03462666604253981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,768,0.00665155549844106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,10240,0.18273777431911895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,512,0.08118489053514269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,10240,0.01184266640080346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,768,0.0038328886859946777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,64,0.0032951111594835916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,2048,0.010004444254769219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,1536,0.004882666799757215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,65536,8.174429999457464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,512,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,768,0.004435555388530095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,65536,5.159923553466797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,7168,0.020149333609475028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3584,1024,0.05703910854127672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,3072,0.10204622480604385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,16384,0.04875377813975016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,1024,0.005597333527273602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,10240,0.8056257565816244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,16384,0.049076444572872586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,5120,0.010199999643696679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,512,0.006871999965773688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,32,0.005576000031497743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,2560,0.13575022750430638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,2048,0.044495999813079834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,6144,12288,0.05552000138494703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,512,0.004643555730581284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,512,8192,0.01260088880856832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,768,0.0058453331391016645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,6144,0.1716195609834459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,128,0.01091466678513421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,2560,0.11952443917592366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,2048,0.033795555432637535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,3584,0.05535644292831421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,5120,0.03242399957444932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,64,0.01033333357837465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,2560,0.012668444050682915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,8192,0.022026666336589392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3584,51200,0.10336977905697292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,3584,0.15438132815890843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,3584,0.018486221631368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,3072,0.009352000223265754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,64,0.008975110948085785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,5120,0.03519644339879354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,5120,0.015400000744395785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,256,0.01088355564408832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,65536,0.12877866956922743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,2048,0.07061244381798638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,5120,0.033097777101728655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,4096,0.01127466642194324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,256,0.0032319999817344877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,4096,0.014529777897728814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,3584,8192,0.04679822259479099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,6144,0.01257955531279246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,64,0.008827555510732863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,768,0.015190222197108798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,65536,0.01350844403107961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,2560,5120,0.07487289110819499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,32,0.008644444247086843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,768,0.028326223293940227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,3584,0.07212089167700873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,6144,0.10137422217263116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3584,1536,0.007089777953094906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,10240,0.01667644414636824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,256,0.012420444852775998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,5120,0.006295111030340195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,1536,0.04356444544262356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,16384,0.023033777872721355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,10240,128,0.01813155578242408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,2560,0.1020124422179328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,3072,0.00741244438621733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,768,0.0040835556056764396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,7168,0.025469332933425903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,2560,0.11774578359391953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,5120,0.048918220731947154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,7168,0.02937955657641093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,3072,0.027313777142100867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,64,0.0029084444459941653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,7168,0.006521777974234686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,7168,4.3450063069661455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,32,0.011555555793974133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,1024,0.04038044479158189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,16384,0.07494755585988362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,256,0.01404088901148902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,2560,0.05357244279649523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,64,0.006678222368160884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,1024,0.022110222114457026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,6144,1.4836435317993164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,512,0.11132000552283393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,1024,0.007456888755162557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,1536,0.01012711144155926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,32,0.013948443863126965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,10240,0.025674665967623394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,7168,0.013618666264745923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,32,0.0036195553839206696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,51200,1.292127079433865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,3584,0.010643555886215635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,768,0.006720888945791457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,4096,0.21020355489518908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,12288,0.025732444392310247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,5120,0.016726222303178575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,12288,0.021350221501456365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,65536,768,0.043947554296917386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,65536,10240,0.33471377690633136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,3584,0.02765688962406582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,3072,0.021336888273557026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,51200,0.0638186666700575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,51200,0.30126934581332737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,16384,3584,0.023842665884229872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,7168,0.04105066590838962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,256,0.01901333365175459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,5120,0.010781333678298526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,512,768,0.003773333297835456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,6144,0.09734488858116998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,4096,0.009630222287442949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,5120,0.01440355512830946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,5120,0.014711999230914645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,1024,0.0058142224119769196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,256,0.0069573331210348345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,2048,0.056758224964141846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,65536,0.09373688697814941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,6144,0.06823199987411499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,2048,0.007151111132568783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,3584,0.09692266914579604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,51200,0.06049066781997681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,16384,0.12176088492075603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,6144,0.02180266711446974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,1536,5120,0.01292888902955585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,3584,0.07374044259389241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,1024,0.006360000206364526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,2560,0.5563057793511285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,768,0.05854666895336575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2048,6144,0.018358222312397428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,768,0.0058444444504049085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,3584,0.11168089177873398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,12288,62.2903578016493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,1536,0.008411555654472774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,1536,0.008105778031879002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,7168,0.06748533248901367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,8192,0.7212364408704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,2048,0.015895111693276297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,1536,0.013674666484196981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,8192,0.016020443704393175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,10240,0.024692444337738886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,32,0.002941333378354708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,4096,0.1423226727379693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,65536,128,0.01034755590889189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,128,0.010575110713640848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,3584,0.017421333326233756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,1024,0.02623644471168518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,512,0.009210666848553551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,3584,0.022237333986494277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,128,0.004643555730581284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,32,0.0028773333049482773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,1536,0.008784888519181145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,16384,0.020344000723626878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,51200,0.9673680199517144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,6144,0.053208887577056885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,768,0.009085333181752099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,512,0.0053377776510185665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,32,0.005084444251325396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,1024,0.3517137898339166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,64,0.008645333349704742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,51200,0.35298400455051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,6144,0.045682665374543935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,256,0.00300799993177255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,3584,0.11055644353230794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,256,0.03152889013290405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,2048,0.007442666424645319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,4096,0.01761777698993683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,10240,0.048024889495637685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,1536,0.06736977895100911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,12288,0.128330667813619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,64,0.00811288919713762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,16384,0.03593777616818746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,1024,0.006168889088763132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,512,0.007772444850868649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,768,0.04043911231888665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,32,0.003208000005947219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,16384,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,16384,0.022132444712850783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,3584,0.09986844327714707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,8192,6144,1.2320044835408528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,2560,0.03933155536651611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,12288,0.014268444644080268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,8192,1536,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,6144,0.10418577988942464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,3584,0.01847822301917606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,3072,0.006254222244024277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,7168,0.013364444176355997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,256,0.02794399857521057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,10240,0.19984711541069877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,1536,0.04031999905904134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,256,0.004202666795916027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,7168,0.011181332998805575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,12288,2.000660366482205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,2048,0.013571555415789286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,2560,0.10578044255574544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,32,0.003126222226354811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,768,0.03867200016975403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,2048,0.018239999810854595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,2560,0.015971556305885315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,1024,0.010046222143703038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,65536,0.10498755507998997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,6144,0.024463999602529738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,7168,0.015234667393896313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,768,0.004899555610285865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2048,51200,0.05827644136216906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,1024,0.006874666859706243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,2560,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,8192,64,0.006055111272467508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1536,128,0.008261333737108443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,1536,0.006256000035338932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,51200,0.07457066906823052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,512,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,768,0.003632889025741153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,16384,0.02276800076166789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,3072,0.008654221892356873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,3584,0.0199635558658176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,10240,0.7489475674099393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,4096,0.016199111938476562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,65536,0.7300951215955945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,768,0.026847999956872728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,4096,0.029120888974931505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,12288,0.032299555010265775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,256,0.006414222220579783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,1536,0.012087110843923358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,3584,0.04725511206520928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,3584,0.008096888661384583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,768,0.006082666830884085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,128,0.004272888931963178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,51200,3.2996383243136935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,12288,1536,0.013367999758985309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,2560,0.06615910927454631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,768,0.006326222171386083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,65536,0.999241828918457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,5120,0.015046222342385186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,65536,13.312010023328993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,7168,0.01995377739270528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,2560,0.008638222184446123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,16384,0.2002711163626777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,12288,0.03234400020705329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,128,0.01314133322901196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,5120,0.03997155692842271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,65536,2.212805430094401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,3072,0.009848000274764167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1536,3584,0.012019555601808759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,64,2048,0.03896444373660617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,5120,0.016551110479566786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,1536,0.010449777874681683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,6144,0.04539111256599426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,3072,0.04076888826158311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,4096,1024,0.00611200018061532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,51200,0.10089155700471665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,16384,0.28876890076531303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,3584,0.03344088792800903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,128,0.0031226666437255014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,512,0.007328000333574083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,5120,0.3377795484330919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,51200,4.183729807535808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,2560,0.01533777846230401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,6144,0.014175999495718213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,16384,0.3186693456437853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,12288,0.14664710892571345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,7168,0.010788444843557147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,3584,0.01703466640578376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,3584,0.014656888114081489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,10240,0.019191111127535503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,768,7168,0.011933333343929715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,2048,0.030768887864218816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,32,0.03271377748913235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,4096,3072,0.09631377458572388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,2560,0.008330666356616551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,4096,0.052058666944503784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,7168,0.09292799896664089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,6144,0.015639111399650574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,128,0.006145777801672618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,3072,0.5650657547844781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,1024,0.006025777922736273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,3072,0.031826668315463595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,512,0.006107555495368109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,7168,0.0370053317811754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,64,0.02461244497034285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,512,0.025369778275489807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,512,0.003535110917356279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,64,0.003104000041882197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,4096,0.02090577781200409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,4096,0.1258044507768419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,5120,0.02383111086156633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,6144,0.014176888598336114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,6144,0.006169777777459886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,1536,0.024647111694018047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,3584,0.02810933192571004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,1024,0.4210960070292155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,3072,0.006052444378534953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,4096,0.033831109603246055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,4096,0.013366222381591797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,512,0.025190222594473097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,10240,0.3855626583099365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,8192,0.01219555570019616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,16384,2048,0.17874577310350206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,16384,1.5998968548244898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,768,1024,0.011611555185582904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,2560,0.008083555433485243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,4096,0.010197333163685268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,8192,0.14870399898952907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,1536,0.6445928679572211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,64,0.006843555718660355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,12288,2048,0.23589777946472168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,128,0.027497778336207073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,512,32,0.009008888569143083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,128,0.13675288359324136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,12288,0.03760622276200189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,32,0.005504888792832692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,768,0.007601777712504069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,2560,0.05240266521771749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,1536,0.01865688959757487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,4096,0.013427555561065674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,7168,0.19695288605160186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,4096,0.05901244613859388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,12288,0.02492444382773505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,51200,1.3020470937093098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,65536,0.05965333514743381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,16384,0.30081245634290904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,128,0.003101333354910215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,2560,0.055768887201944985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,65536,1.3841057883368597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,5120,0.033972445461485125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,4096,0.016882666283183627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,2560,0.015170666906568738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,3584,0.023299554983774822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,10240,0.06385955545637342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,4096,0.4007546636793349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,5120,0.02232444451914893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,8192,0.06393866406546699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,51200,3072,2.0816097259521484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,64,0.02663199934694502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,16384,0.17075644599066842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,32,0.06994933552212186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,768,0.004931555440028508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,1024,256,0.02683199942111969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,51200,1.6431218253241644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,65536,0.08917511171764797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,5120,0.5172924465603298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,3072,0.06457866562737359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,51200,1.3348684310913086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,768,0.006279110908508301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,5120,0.15709511439005533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,1024,0.021527111530303955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,2048,0.0120328888297081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,2560,0.01938666734430525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,8192,0.015962666935390897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,6144,0.473723570505778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,1536,0.023363555471102398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,3584,0.16435999340481228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,128,0.003658666792843077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,51200,12288,0.18692978223164877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,64,0.02442755632930332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,64,0.002789333255754577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,1536,0.06714133421579997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,12288,0.07037778033150567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1024,10240,0.015934222274356417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,64,0.013425778183672162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,2560,0.012483555409643384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,1536,0.011230222052998014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,65536,0.056596444712744824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2048,2560,0.008479110896587372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,2048,0.06956533590952556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,5120,0.019315555691719055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,1024,0.050965332322650485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,1024,0.02641511129008399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,32,3584,0.008358221915033128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,512,0.01683111157682207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,512,4096,0.006446222050322427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,65536,8192,4.537545098198785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,256,0.005431111074156231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,1536,0.008096000386608971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,10240,0.023992889457278784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,128,0.005336888962321811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,12288,0.26872534222073025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,512,0.007592889169851939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,128,0.004120888809363048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,7168,0.10988089111116196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,768,0.026781333817376032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,16384,16384,0.9833964241875542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,16384,0.02738933265209198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,64,0.014835556348164877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,32,0.005895111295912001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,65536,0.4566737810770671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,5120,0.058824002742767334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,768,0.005170666509204441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,32,0.0027831111931138565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,64,0.002945777856641345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,128,0.03652088840802511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,1536,0.00693866651919153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,16384,0.544232898288303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,10240,0.2667226791381836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,8192,4096,0.016232000456915963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1024,5120,0.02797422144148085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,16384,1024,0.011780444118711682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,51200,0.8927893108791776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,3584,0.015604444675975375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,12288,0.04034844372007582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,1536,0.09437866343392266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,12288,1.3342569139268663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,2048,0.00742222203148736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,768,0.04458399944835239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,12288,0.018968888454967074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,2048,0.01611911091539595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,2048,0.012310222619109683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,768,16384,0.01236355553070704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,7168,0.23047110769483778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,32,0.0052666668262746595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,3584,0.009679111341635386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,6144,0.009452444811662039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,256,0.019066666563351948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,1536,0.012796444197495779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,256,0.0038382220599386427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,65536,0.14765955342186823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,2048,0.013978666729397245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,512,0.003934222377008862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,16384,0.031552000178231135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,2560,0.007904888855086433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,4096,0.01203733351495531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,6144,0.020523554748959012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,51200,0.2710328896840413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,12288,0.9033092922634549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,65536,0.17632355954911974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,12288,0.023524444964196947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1536,3072,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,1536,0.04581244455443489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,51200,8192,0.45905335744222003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,5120,0.1913599967956543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,12288,0.34353065490722656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,10240,0.062096887164645724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,7168,0.10815555519527859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,128,0.006042666733264923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,128,0.004962666581074397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,6144,0.014311111635631986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,12288,0.014560000763999091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,2048,0.04083022144105699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,4096,0.027186666925748188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,64,0.02905155552758111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,51200,19.99722629123264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,2560,0.009310222334331935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,512,0.006053333481152852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,3072,0.023783110909991797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,1024,0.005943999936183293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,16384,0.06460177898406982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,1024,0.05994133154551188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,16384,0.4316506650712755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,6144,0.015805333852767944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,16384,3072,0.019695111446910434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,51200,0.05942844682269626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,6144,0.015828443898095023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,1536,0.0704079998864068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,6144,0.031095110707812842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,128,0.028431110911899145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,8192,0.036779556009504534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,51200,32,0.11455467012193467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,51200,0.11321155230204265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,7168,0.3959662119547526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,64,0.008864888714419471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,128,0.020169778002632987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,2048,0.415791114171346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,10240,0.03168177935812209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,3072,0.021168000168270532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,128,0.018177777528762817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,8192,1.1721466912163627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,256,0.00508088908261723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,512,1536,0.0069368887278768755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,16384,0.06320977873272367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,32,0.00648266656531228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,256,0.004062222109900581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,768,2048,0.027404444085227117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,768,0.006966222491529252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,6144,0.015957333975368075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,6144,0.6829839812384711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,768,0.005502222312821283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,65536,0.07470577955245972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,1536,0.15527821911705866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,10240,0.007859555383523306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,5120,0.026092444856961567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,10240,0.026554667287402686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,512,4096,0.009080000221729279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,12288,0.006144000010357962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2048,3584,0.011155555645624796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,64,0.11650933159722222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,65536,0.36141334639655215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,1024,0.004577777865860197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,5120,0.4536302354600694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,256,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,16384,2.9947217305501304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,1024,0.008974221845467886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,7168,2.53841421339247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,10240,0.025591111845440332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,8192,0.017901332841979135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,3072,0.1889582210116916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,65536,0.14527111583285862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,64,0.013564444250530668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,768,0.036497778362698026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,10240,0.06319200330310397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,2560,0.01519733336236742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,1024,0.025944888591766357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,8192,0.006441777778996362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,4096,0.02778844369782342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,1024,0.0069653333889113525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,16384,0.5231893327501085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,5120,8192,0.01981244484583537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,12288,12288,0.40834487809075254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,4096,0.04706133405367533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,51200,0.015818667080667283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,64,0.008067555725574493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,4096,0.21943733427259657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,7168,0.006365333166387346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,768,0.0042257776690853965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,8192,10240,0.39819733301798504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,2560,0.01017155581050449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,7168,0.019517333971129525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,12288,0.02846755584081014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,2048,0.05380622214741177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,6144,0.022506665852334764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,10240,0.03815199931462606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,64,0.002728888971938027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,16384,0.17078044679429796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,256,0.011583111352390714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,16384,0.36717955271402997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2560,2560,0.007613333563009898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,512,0.005482666608360078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,10240,0.01734755602147844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,2560,0.7671164406670464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,16384,0.02832088867823283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,12288,0.02648711038960351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,768,0.03515999846988254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,10240,0.4672364658779568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,65536,5.49365954928928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,4096,0.02162133322821723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,3584,0.23151911629570854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,128,5120,0.009544889132181803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,3584,0.10956533087624444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,1536,0.006393777827421824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,65536,0.9594222174750434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,2560,0.10726933346854316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,1536,0.004834666848182678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,1536,0.006298666612969504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,51200,2.6164364284939237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,1536,0.007163555257850223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,512,0.06757599777645536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,512,0.03692355420854356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,1536,0.005075555708673266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,4096,0.010128888818952773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,128,0.012152888708644442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,32,128,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,768,0.00618488879667388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,3072,0.01885955532391866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,7168,0.16798933347066244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,12288,0.050755557086732656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,10240,0.04091466797722711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,1536,0.007032000356250339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,3584,0.010159999959998662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2560,65536,0.10065688689549764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,65536,0.40498489803738064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,16384,0.018552889426549275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,16384,51200,21.422125922309025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,5120,0.1009555525249905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,2560,0.008020444048775567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,65536,0.2154231071472168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,128,0.0034337778472238113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,4096,0.01033333357837465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,256,0.008151999778217739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,768,0.0047306666771570844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,5120,0.006969778074158563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,7168,0.02414400047726101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,2048,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,16384,0.09663199716144139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,768,0.0038168889780839286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,65536,0.10971199803882176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,5120,3.2210036383734812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,65536,16.78173658582899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,6144,0.017352888981501263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,512,0.006765333314736684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,3584,0.05726933479309082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,1536,0.014486221803559197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,768,65536,0.04682666725582547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,8192,1024,0.010804444551467896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,3072,0.008234666453467475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,8192,0.01665155589580536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,6144,0.17931021584404838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,10240,256,0.007463999920421177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,16384,0.09682666593127781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,128,0.03901422354910109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,128,6144,0.010205333431561789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,10240,0.052779555320739746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,32,0.012379555238617791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,3584,0.007930667036109501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,3072,0.009036444127559662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,1536,0.010656000011497073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3072,51200,0.09683911005655925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,256,0.00817599975400501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,10240,0.08025778002209134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,4096,0.09178666936026679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,3072,0.03783555494414435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,10240,0.013658666776286231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,16384,0.07338311274846394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,1024,0.02792444494035509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,1536,0.006050666587220297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,3072,0.022113778524928625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,512,0.011411555939250521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,256,0.003067555526892344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,32,51200,0.4205191135406494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,65536,0.9032444424099393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,3072,0.023354666100607976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,51200,0.9640710618760852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,12288,2.160034603542752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,1024,0.02418577836619483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,4096,0.03103289008140564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,64,0.009476444787449306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,7168,0.019299555155966017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,512,0.007113777928882175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,3072,0.012278222375445895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,128,0.0061048890153567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,6144,12288,0.5981813536749946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,1024,0.1566186613506741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,1536,0.0724124444855584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,512,10240,0.009024000002278222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,1024,0.005488889084921942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,6144,0.1618222263124254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,64,0.022038223014937505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2048,64,0.005480888817045424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,7168,0.16069511572519937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,256,0.0069848886794514125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,768,7168,0.011557333171367645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,65536,10240,0.3318355613284641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,51200,0.010310221877362993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,3072,0.017296888762050204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,65536,5.5207917955186625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,16384,0.079184889793396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,32,0.0029235554652081597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,128,0.11668799983130561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,5120,0.03283377819591098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,6144,10240,0.02679022153218587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,512,0.006708444406588872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,8192,0.1180008914735582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,16384,0.05238222082455953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,12288,0.06795910994211833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,768,0.009773333039548662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,3072,0.0965119997660319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,1536,0.034064001507229276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,32,0.04636089007059733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,3584,0.2744968997107612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,1536,0.004815111143721474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,12288,0.01051999959680769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,8192,0.087281776799096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,1536,0.010618666807810465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,65536,0.08877599901623195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,5120,0.12977688842349583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,3072,0.15350488821665445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,64,0.008871999879678091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,2048,0.020417778028382193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,32,0.0025048889219760895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,768,0.004825777891609403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2048,2048,0.02271555529700385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,10240,3.344886144002279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,2560,0.008543999658690559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,8192,0.25910043716430664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,7168,0.00887022250228458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,16384,0.610873752170139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,128,0.003342222215400802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,768,0.029503110382292006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,3584,0.09479466411802505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,1024,0.011649777491887411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,3584,0.010724444356229572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,16384,0.05156888895564609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,65536,0.12160089280870225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,256,10240,0.01645599967903561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,4096,0.2041084501478407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,4096,0.17543200651804605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,256,0.004851555658711327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,256,0.011351111034552256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,2560,0.00832177781396442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,10240,0.041839109526740186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,256,1536,0.0074684446056683855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,1536,0.09161688884099324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,51200,0.8660880194769965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,8192,0.03175910976197984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,64,2048,0.007148444652557373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,2048,0.007727999654081132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,256,7168,0.01165688865714603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,51200,1.0761431588066948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,3584,0.018107554978794523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,2560,0.026848889059490625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,10240,0.022245332598686218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1536,7168,0.04963111215167575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,5120,0.006815111057625876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,7168,0.007134222322040134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,51200,0.45384624269273544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,2048,0.01535555554760827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,6144,0.041858666472964816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,3584,0.11246489153967963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,3072,0.05583110782835218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,65536,0.4008791181776259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,51200,0.05236177643140157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,512,0.037463112009896174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,2560,0.8199849128723145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,5120,1.0225458145141602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,6144,0.03763822383350796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,128,0.005673777725961473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,768,0.009326222042242685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,1536,0.00904622260067198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3072,8192,0.178077326880561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,12288,768,0.009914666414260864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,4096,0.06863466898600261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,51200,3.961210462782118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,51200,0.08200355370839436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,12288,0.5796213150024414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,32,0.0028773333049482773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,51200,2.904255972968208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,7168,0.007875555919276344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,8192,0.668726232316759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,128,0.015736889508035447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,5120,0.03014311194419861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,2048,0.09344977802700466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,3584,0.010588444769382477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,65536,0.015406222807036506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,32,0.03846044341723124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,12288,0.727110226949056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,51200,1024,0.03258400162061056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,65536,1.8458105723063152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,1024,0.1204106675253974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2048,12288,0.1917733351389567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,10240,0.02331999937693278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,64,0.0029066666546795103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,51200,0.5375368860032823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,65536,1.5410595999823675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,3584,0.021799110703998145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,12288,2.2805697123209634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,6144,0.01370666672786077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,12288,0.05332444442643059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,64,0.0074275558193524676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,32,0.0035751110149754416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,1536,0.10311733351813422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,256,2048,0.0074275558193524676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,16384,0.05034666591220432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,2560,0.009269333548016017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,4096,0.012727999852763282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,2048,0.0053093334039052325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,128,0.009893333746327294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,12288,0.03250577714708116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,768,0.0058000000814596815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,8192,0.01624000072479248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,512,0.00573333352804184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,5120,0.014188443620999655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,1024,0.006071110980378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,2560,0.010727110836240979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,1536,0.04325422313478258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,16384,0.01930488811598884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,4096,3072,0.013088888592190213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,512,0.0035457776652442086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,256,0.003637333297067218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,3584,0.17595911026000977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,10240,768,0.009199110998047723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,768,0.016920889417330425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,256,0.003083555648724238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,51200,8.537307739257812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,1024,0.034605334202448525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,5120,0.01702933344576094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,10240,0.18703822294871011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,10240,0.7380391226874458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,128,0.014341332846217685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,5120,0.02180266711446974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,512,0.004974222017659081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,2048,0.00646666685740153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,5120,16384,0.3731493420071072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,16384,4096,0.022447110878096685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,1024,0.006505777852402792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,1536,0.0069546666410234236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,256,0.06986222002241346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,12288,0.15125866731007895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,768,128,0.004795555439260271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,1024,0.04421155651410421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,16384,8.00651380750868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,256,0.004089777668317159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,3072,0.07092177867889404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,16384,0.028170665105183918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,12288,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,8192,0.04033777780003018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,2560,0.010672888822025724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,3072,0.015251555376582675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,1536,0.009516444471147325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,16384,0.17988177140553793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,6144,2.7297787136501737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,256,3584,0.02535555594497257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,7168,0.14951022466023764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,65536,0.6277813381618923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,128,0.005610666755172942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,256,0.023760888311598036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,1536,0.014745778507656522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,12288,64,0.035699556271235146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,2560,0.05112089051140679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,1024,0.0459537771013048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1536,64,0.008488000267081791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,7168,0.013427555561065674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,64,0.0033288889875014624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,128,0.0034640000926123727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,65536,3.8787189059787326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,1024,0.005611555443869696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,1536,0.007680888805124495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,128,0.01053422192732493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2560,65536,0.6715848710801867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,6144,3.7354647318522134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,5120,0.2189075549443563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,16384,0.007286222444640265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,5120,0.02259911100069682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,2048,0.009682666924264696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,2560,0.007137777904669444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,16384,0.04121511181195577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,512,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,3072,0.01182755579551061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,2048,0.5446524620056152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,768,0.0653466648525662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,1536,0.00609777785009808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,6144,10240,0.44759379492865664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,51200,0.2607733408610026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,512,0.026683555708991155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,16384,0.8440675735473633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,4096,10240,0.03186577889654372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,10240,0.031699554787741765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,32,0.010591111249393888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,6144,0.01939288940694597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,128,0.005567111074924469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,1536,0.0243848893377516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,5120,0.2258417871263292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,32,0.022078222698635522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,256,0.15977511141035292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,2048,0.012011555333932241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,65536,20.612838745117188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,2048,0.04840622345606486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,512,0.009128000173303816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,1536,0.01260088880856832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,5120,0.008420444197124904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,32,0.003137777869900068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,3072,0.008574222524960836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,4096,0.5721919801500108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,1024,0.027935110860400732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,128,0.012797333300113678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,3584,0.007894222107198503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,32,0.01017599966790941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,768,0.006721777634488211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,32,0.011591110792424945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,5120,0.04244355691803826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,16384,0.024289778537220422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,768,0.006263999889294307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,51200,1536,0.5109635459052192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,768,0.0144533332851198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,8192,0.01423733267519209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,7168,0.14085511366526285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,7168,1.3871964348687067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,1024,0.2524675528208415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,51200,0.4652133517795139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,3072,0.008107555409272512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,32,0.0027217778066794076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,2560,0.007556444240940942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,2560,0.014534221755133735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,12288,0.017197334104114108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,256,0.015825778245925903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,3072,0.10525777604844834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,2048,0.006888888776302338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1536,2560,0.009020444419648912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,65536,0.0693368911743164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,6144,0.5816355811225044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,1024,0.026547556122144062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,12288,0.034950223233964704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,8192,0.03275377882851495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,1024,0.04234222239918179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,6144,7168,0.10931022299660577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,16384,0.05234488844871521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,3072,0.006251555350091722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,16384,0.05883200301064385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,256,51200,0.3937697675493028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,6144,4.42629157172309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,65536,0.11632000075446235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,6144,0.017910222212473553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,3584,1.719646241929796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,128,0.003216000066863166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,768,0.02215911116864946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,2560,0.009798222117953831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,65536,0.132168001598782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,2560,0.008563555777072906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,12288,0.018031110366185505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,10240,0.02614044480853611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,5120,0.06327111191219754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,1024,0.004281777888536453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,3584,0.00742222203148736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,8192,2048,0.011027555498811932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,512,0.00387999994887246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,32,0.028263999356163874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,8192,0.015488000379668342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,7168,1.7012471093071833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,1536,0.007823111282454597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,1024,0.036546667416890465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,128,0.008099555969238281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,4096,0.007979555262459649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,32,0.012135110795497894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,51200,0.22637332810295951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,2048,0.013859555953078799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,768,0.019173334042231243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,7168,1024,0.029149333635965984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,4096,0.3917422294616699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,6144,0.1861013306511773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,512,0.0033928888539473214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,256,0.014257777068350049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,768,0.012471111284361945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,6144,0.012850667039553324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,65536,16384,7.3906809488932295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,8192,0.2513848940531413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,32,0.005410666680998272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,65536,0.44123021761576336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,2048,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,10240,6144,0.4237324396769206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,768,0.028283556302388508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,3584,0.04774844315316942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,1024,0.004449777719047335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,3584,0.024145777026812237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,7168,0.023269333773189124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,51200,0.5462275611029731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,51200,0.23787911732991537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,512,0.005671111245950063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,5120,0.04958577619658577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,51200,0.05554044577810499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,65536,0.04822311136457655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,128,0.004411555412742826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,12288,0.03355288836691115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,1024,0.006061333335108227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,2048,0.03344355689154731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,8192,0.1561946603986952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,12288,0.9786844253540039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,3584,0.017501332693629794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,128,0.004243555582231945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,64,0.004377777791685528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,2560,0.007599111232492659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,5120,0.45835553275214297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,1024,0.009758222434255812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,12288,0.017609778377744887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,6144,0.43264089690314395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,32,0.005807111246718301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,64,0.002982222164670626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,3072,0.010149333212110732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,7168,0.011925333076053195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,3072,0.026226666238572862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,7168,0.05570133527119955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,6144,2.1221546596950955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,65536,0.48723199632432723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,64,0.006324444380071428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,3072,0.007327111230956183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,512,0.004548444516128964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,1536,0.009045333498054082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3072,6144,0.0185199992524253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,16384,0.058614220884111196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,5120,0.024077332682079736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,1536,0.006110222389300664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,256,0.02218399941921234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,64,0.020224889119466145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,128,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,65536,2.7008391486273875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,32,0.016900445024172466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,32,0.00958577791849772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,2560,0.03151911165979173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,3584,0.02848177817132738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,768,0.025719111164410908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,256,8192,0.02463111115826501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2560,64,0.006481777876615524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,3072,0.1575102276272244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,65536,0.03056000007523431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,3072,0.013619555367363824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,6144,0.03622044457329644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,16384,0.026978666583697002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,256,0.006214222146405115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,6144,0.01866222255759769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,5120,0.02959200077586704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,65536,0.5507111019558376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,8192,0.13470044400956896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,6144,0.014025777578353882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,256,0.02681066592534383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,1536,0.0064035554726918536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,6144,0.019800000720553927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,768,1536,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,2048,0.01744888888465034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,3584,0.06684533092710707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,1536,0.004798222333192825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,65536,512,0.021687111920780603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,65536,3072,2.647523456149631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,2048,0.06261866622500949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,8192,0.08745155731836955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,10240,7168,0.05714844332800972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,16384,768,0.010487111078368293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,2048,0.007152888509962294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,16384,5120,0.07597333192825317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,10240,0.009637333452701569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,512,0.030618667602539062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,16384,0.1288879977332221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,768,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,5120,0.011952888634469775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,12288,0.4530337651570638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,1024,0.025148444705539282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,3584,0.033717334270477295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,64,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,128,0.005148444324731827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,3072,0.0063697778516345555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,10240,0.023214222656355962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,512,0.005576000031497743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,7168,0.7224959797329373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,6144,0.024496000674035814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1536,5120,0.03429155548413595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,6144,0.031360887818866305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,128,0.00286666676402092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,51200,0.5588453080919054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,1024,0.07232889201905993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,3584,0.20868355698055693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,256,0.022864888111750286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,10240,0.849822203318278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,2560,0.020057777563730877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,10240,0.02830933199988471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,768,0.006825777805513806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,12288,0.018961777289708454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,10240,0.02377066678471035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,65536,0.07832533121109009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,4096,0.049802667564815946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1536,16384,0.44001065360175234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,12288,0.21974399354722765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,3584,0.12098488542768691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,5120,2560,0.09043288893169826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,1024,0.006280888699822956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,12288,0.007583111524581909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,4096,0.046282665597067944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,1536,0.006600889066855113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,65536,3.134233686659071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2560,3072,0.015593777100245157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,7168,4.240230984157987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,768,0.007155555817815993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,512,0.016196444630622864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,128,0.0041333333485656315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,3072,0.035601778162850276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,128,0.006218666831652324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,16384,128,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,2048,0.00666133314371109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,51200,0.44574753443400067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,512,0.010519111322032081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,4096,0.030608889129426744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,12288,0.019222221440739103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,512,0.004354666504595014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,10240,0.014631110760900708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,2048,0.05968177980846829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,768,0.14535644319322374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,65536,1.6239564683702257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,768,0.008479110896587372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,6144,0.028247998820410833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,1024,0.04823199907938639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,8192,0.0382008883688185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,64,0.0068986668354935115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,1024,0.007833778030342525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,256,0.0034355556385384668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,32,0.0025946667624844448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,7168,0.05456533365779453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,3584,0.016220443778567843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,65536,0.683935112423367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,2048,0.005671999934646819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,1024,0.007228444847795699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,4096,2048,0.046726223495271474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,32,0.005335999859703912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,5120,0.015988444288571674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,3584,0.02514133354028066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,10240,0.024293333292007446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,7168,12288,0.03158488869667053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,512,0.02514222264289856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,6144,0.18299378289116752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,5120,0.012585777375433179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,65536,0.06376088990105523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,12288,0.049434665176603526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,4096,0.011482666763994427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,7168,0.12587732738918728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,16384,768,0.04490844408671061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,3584,0.012181333369678922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,768,0.005411555369695027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,2560,0.1320319970448812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,1536,0.013978666729397245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,16384,0.020918221937285528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,16384,0.021335111724005804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,10240,0.023151111271646287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,51200,0.09788978099822998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,512,0.005326222214433882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,32,3072,0.00794400026400884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,51200,0.09919910960727268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,65536,3.4937911563449435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,768,0.043750223186280995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,128,0.006960888703664144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,16384,0.19547643926408556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,6144,0.04907288816240099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,256,0.0028764444092909494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,6144,0.015375999940766228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,256,0.019464888506465487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,6144,0.11722044150034587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,5120,0.06527466906441583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,512,0.00452888881166776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,65536,7168,0.26162489255269367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,65536,8.279433356391058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,16384,0.025839999318122864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,12288,0.01519466605451372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,768,0.01020088874631458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,128,0.006244444598754247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,2560,0.040911999013688825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,3072,768,0.012524444195959302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,16384,16384,1.6831697887844508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,256,0.01736088924937778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,1536,0.022487999664412603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,128,0.008959111240175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,1536,0.009478222164842818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,51200,1.1855715645684135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,5120,0.1064515577422248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1024,5120,0.011527111132939657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,65536,5.516440073649089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,3584,0.011809777882364062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,32,0.008200888832410177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,8192,0.03983733389112685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,256,0.012718222207493253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,8192,65536,3.9750879075792103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,128,0.008184000021881526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,256,0.012254222399658628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,64,0.007142222589916653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,10240,0.015127999915017022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,2560,0.0314337776766883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,768,3584,0.028539554940329656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,1536,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,8192,0.022395556171735127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,65536,1.3160009384155273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,16384,0.009954666925801171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,3072,0.019552888141738046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,8192,0.3027084403567844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,7168,0.006656000183688269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,64,768,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,8192,0.028703110085593328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,5120,0.00981955561372969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,512,0.009310222334331935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,4096,0.04699644446372986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,128,0.11019200086593628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,256,0.0031866667171319327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,10240,0.05926666657129923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,1024,0.004943999979231092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,64,768,0.009296000003814697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,3584,0.014853333433469137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,10240,0.026684444811609056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,1536,0.006531555619504716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,12288,0.026703110999531213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,6144,0.03906311260329352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,3072,0.013945778210957846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2560,2048,0.0650764438841078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,768,0.019111111760139465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,256,0.011615110768212212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,3584,0.021047110358874004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,3584,0.11679022841983372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,64,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,768,0.0037875556283526947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,32,0.013839999834696451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,32,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,51200,1.212930679321289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,10240,0.4866613282097711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,5120,0.13477778434753418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,128,12288,0.019807111885812547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,32,0.0034968890249729156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,65536,0.23366665840148926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,6144,0.102528002527025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,5120,0.3890053431193034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,4096,1536,0.04169688953293694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,2560,0.00923822240697013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,2560,0.025800888737042744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3072,256,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,768,0.012545777691735161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,2560,0.009015999734401703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,4096,0.040885332557890154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,4096,0.17573066552480063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,2048,0.010515555739402771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,768,4096,0.011501332951916588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,1536,0.05252888798713684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,3584,0.18394488758511016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,51200,0.9873093499077691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,2048,0.05752266777886284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,1024,0.01587288909488254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,12288,32,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,1536,0.097543113761478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,64,0.0033599999215867785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,16384,0.7548089027404785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,12288,3072,0.02071377800570594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,3072,0.021264889174037512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,1024,0.03924177752600776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,1536,0.006312888943486744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,1024,0.0072266666425599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,4096,0.016474665866957772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,8192,64,0.01001422190003925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,1024,0.0058133333093590206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,768,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,8192,0.17664266957177055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2560,4096,0.015746666325463187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,65536,1.4491839938693578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,2048,0.008302222523424361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,3072,0.14494400554233128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,1536,0.0063786668082078295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,64,0.0026871110830042097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,3584,0.02316266629430983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,7168,0.011200000014570026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,32,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,6144,0.010263111028406356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,512,0.00720977783203125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,10240,0.03267644511328803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,65536,0.009101333717505137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,3584,0.00693866651919153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,6144,0.012464889221721225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,3584,768,0.0329751107427809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,512,8192,0.009832888841629028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,1024,0.010948444406191507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,768,0.016599110431141324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,3072,0.010040000081062317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,3072,0.10170577632056342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,12288,0.0650213360786438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,8192,2.6705813937717013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,51200,15.75098843044705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,7168,0.018265777164035372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,1024,0.006175999840100606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,3584,0.02534577747186025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,16384,3.5012764400906033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,6144,0.9313724305894641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,32,0.010784889260927836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3584,2048,0.011634666886594562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,32,0.006068444500366847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,16384,12288,0.5926595793830024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,768,0.0088764445649253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,5120,0.09134133656819661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,128,0.002826666666401757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,4096,4096,0.13755022154914007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,51200,0.18720089064704049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,768,0.6497493320041233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,128,0.025249777568711176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,4096,0.01221333361334271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,7168,0.013293333351612091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,256,0.006904888898134232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1024,5120,0.012140444583363004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,16384,0.03780711028310988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,1536,0.03130311104986403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,7168,0.013607111242082385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,6144,0.01780533293883006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,1024,0.07769866784413655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,2048,0.006350222147173352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,65536,0.03713688916630215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,2048,0.012454222473833295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,512,0.0032960000551409195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,8192,0.015806222955385845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,4096,0.17765066358778211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,5120,0.27380622757805717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,51200,2.087242762247721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,51200,0.013189333180586496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,3584,0.009461333354314169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,8192,0.10221244229210748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,2560,0.2458862198723687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,1024,0.008772444393899705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,51200,10240,0.2666417757670085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,256,0.003272888975010978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,8192,0.569959110683865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,12288,0.4842284520467122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,3584,0.024600888291994732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,512,0.004979555391603046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,512,0.02569244470861223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,2048,0.6647324562072754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,65536,0.0225262227985594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,65536,0.8575359980265299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,8192,0.05222488774193657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,8192,0.02814133299721612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,3072,0.11349333657158746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,65536,0.07354933023452759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,2048,0.007072889142566257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,32,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,1536,0.017302221722073026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,1024,0.006055999961164262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,65536,0.06574844651752047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,768,0.012693333129088083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,8192,0.453895992702908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,1536,0.008849778109126622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,6144,0.05203288793563843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,7168,0.013229333692126803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,512,0.0063742221229606206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,5120,0.16731022463904488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,768,0.004096888833575779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,12288,0.11055911249584621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1024,2048,0.011126221881972419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,512,2560,0.007088888850477006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,2560,0.07450755437215169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,768,2048,0.02297244469324748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,768,0.003601777884695265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,128,0.0038613333470291565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,8192,0.016538666354285348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,512,0.004478222380081813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,1024,0.006146666490369373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,8192,0.03453866640726725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,16384,0.3070053259531657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,3072,0.00849422232972251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,512,0.007623111208279927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,1024,0.028060442871517603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,51200,0.32233066029018825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,768,0.0037715555065208008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,10240,2560,0.1621511114968194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,3072,0.3044782214694553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,5120,0.05570399761199951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,16384,0.023528888821601868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1024,512,0.003546666767862108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,6144,0.29920532968309194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,128,3584,0.010199999643696679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,64,0.019208000765906442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,5120,0.036291556225882635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,2560,0.07250933514700995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,512,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,3584,0.24601422415839302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,51200,0.1564799944559733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,51200,0.1765200032128228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,5120,0.1799644496705797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,51200,0.14336000548468694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,16384,0.026896889011065166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,6144,0.15462756156921387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,1024,0.004062222109900581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,128,0.1209377712673611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,1536,0.08074311415354411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,65536,0.27525777286953396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,2560,0.01719288859102461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,4096,0.010447111394670276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,8192,0.023175999522209167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,768,0.09893866380055745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,12288,2.177984873453776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,7168,0.04256711072391934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,6144,0.10618311166763306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,51200,0.40954134199354386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,51200,0.44529422124226886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,1024,0.01538133290078905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,3584,0.010507555471526252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,32,12288,0.006127111199829314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,64,768,0.008966222405433655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,64,0.0027075554761621687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,51200,0.26647289594014484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,768,0.005097777893145879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,5120,0.010658666491508484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,4096,0.3593182298872206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,5120,1024,0.008089777496125963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,256,0.00960711141427358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,10240,0.08619733651479085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,65536,2.045045428805881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,3072,0.026339555780092876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,65536,0.06658844153086345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,7168,0.08683911297056411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,51200,0.08908177746666802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,6144,0.010606221854686737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,1536,0.03259999884499444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,768,0.0038488888078265717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,2560,0.02162133322821723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,256,0.015187554889255099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,64,0.005087999833954705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,16384,32,0.004990222139490975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,12288,0.05680177609125773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,5120,12288,0.03271289004219903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,7168,0.08209777540630765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,4096,0.010290666586822933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,1536,0.008642666869693333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,6144,10240,0.022837332553333704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,6144,256,0.01771288944615258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,2560,0.11923200554317898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,4096,0.011380444798204633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,10240,1536,0.015449777245521545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,2560,0.021907554732428655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,1024,0.004377777791685528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,512,0.017927999297777813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,10240,51200,0.4525449011060927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,3584,0.010979555547237396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,256,0.006439111298984951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,768,0.01963733302222358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,768,0.03880711065398322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,65536,1.5704622268676758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,128,0.12572533554501003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,10240,0.3247902128431532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,6144,0.016142222616407607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,3584,2560,0.07457333140903048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,4096,0.15425688690609402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,1536,0.008190222084522247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,51200,4096,0.46300797992282444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3072,65536,0.10244711240132649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,16384,4.761785719129774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,6144,6144,1.776975949605306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,12288,0.0125928885406918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,6144,0.020826665891541373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,10240,768,0.04527022110091316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,6144,0.008240000241332585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,51200,1.2982666227552626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,64,0.013984888792037964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,65536,1.3260889053344727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,3584,0.050795555114746094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,2560,0.028156446086035833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,2560,0.027259555127885606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,7168,0.01834755473666721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,6144,0.10470577743318345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,7168,0.11869866318172878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,2560,0.06875110997094049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,4096,0.014292443792025248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,12288,0.0962302221192254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,3072,0.0067528887755341
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,4096,0.1915342277950711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,8192,0.011960888902346293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,768,512,0.003411555455790626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,64,0.007190222541491191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,1024,0.08863555722766453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,6144,0.014467555615637036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,2560,0.11432445049285889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,8192,0.06396444638570149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,256,0.013501333693663279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,768,0.04482577906714546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,7168,6144,0.13868978288438585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,32,0.02104088829623328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,64,0.021689777572949726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,1536,768,0.020618667205174763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,6144,0.48569509718153214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,1024,0.005879111174080107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,2048,0.01310755560795466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,2560,0.014621333943472968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,51200,0.06945511367585924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,2560,0.1453688939412435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,8192,0.050236443678538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,768,0.007803555164072249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,1024,0.005994666781690385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,512,0.22600444157918295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,4096,64,0.013981334037250943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1024,2560,0.045978665351867676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,10240,0.32087379031711155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,768,128,0.005544888890451855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,12288,0.030266665750079687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,1024,0.03945866558286879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,768,0.0273982220225864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,1024,0.006724444528420766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,512,0.35514399740431046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,12288,0.10247021913528442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,12288,0.07018666797214083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,51200,3584,0.05700711409250895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,10240,0.9325804180569119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,10240,0.023936000135209825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,1024,0.04749422272046407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,3072,0.10936799976560806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,7168,0.23448888460795084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,3072,0.058131555716196694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,3584,0.07387733459472656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,768,256,0.01001422190003925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,64,8192,0.006126222097211414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,6144,0.01854044364558326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,6144,0.02276266614596049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,3584,0.1348666614956326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,10240,0.13596888383229574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,1536,0.008839111361238692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,4096,0.023502222365803186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,6144,0.038155555725097656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1536,2560,0.023141332798533972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,65536,0.9531244701809354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,256,0.012488889197508493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1536,7168,0.47804890738593203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,10240,0.6715111202663846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,1536,0.014053333136770459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,2048,0.0685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,8192,0.023409777217441138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,12288,1.161818716261122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,512,128,0.006640888750553131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,12288,0.031026668018764917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,7168,0.01309511148267322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,768,0.006709333509206772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,512,0.005723555468850666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,65536,0.6001564661661783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,32,0.02686311139000787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,2560,0.0068853331936730286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,10240,0.035994665490256414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,1536,0.014334221680959066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,3072,0.01350488927629259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,3584,0.018225777480337355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,3584,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,65536,0.09670933087666829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,3584,0.00852177788813909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,65536,3072,0.4316177897983127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,12288,0.019162666466501024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,8192,0.842708428700765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,65536,8.415257771809896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,4096,0.01963200006220076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,6144,0.018794667389657762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,128,0.0058764442801475525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,512,0.00684533350997501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,10240,6144,0.4035920037163629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,3072,0.026621333426899377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,16384,0.1384062237209744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,256,51200,0.16032800409528944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,3584,0.00796622203456031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,3584,0.11604089207119411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,128,4096,0.021306667062971327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,3072,0.006523555351628198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,512,0.01463822192615933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,768,0.005347555710209741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,3584,0.016905777984195285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,4096,0.44385067621866864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3584,2560,0.009429333110650381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,128,0.0042257776690853965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,2560,768,0.023994666006830003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,2560,0.012530666258600024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,16384,3.999413384331597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,6144,0.027191110783153113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,51200,0.7480115360683866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,2560,0.01367377738157908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,7168,0.01620444489849938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,4096,0.015710221396552194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,512,0.0041555555330382455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,3584,0.06036000119315254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,64,0.006001777946949005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,51200,0.4162471029493544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,51200,0.21344266997443306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,3072,0.01607022186120351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,64,0.015480000111791821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,51200,2.0890543195936417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,6144,0.10511733425988092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,2560,0.04090311129887899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,512,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,10240,0.016889777448442247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,512,0.005630222045713001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,128,0.003863111138343811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,12288,1.5819404390123155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,1024,0.6292391353183323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,1024,0.3338373237186008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,10240,0.01629599928855896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,2048,0.036397334602144026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,2048,0.11214133103688557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,16384,0.052370667457580566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,8192,2.8825626373291016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,2048,0.015376889043384127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,2048,0.008388444781303406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,6144,0.19292267163594565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,10240,0.012984000146389008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,10240,0.013272888958454132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,3584,0.016167110866970487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,6144,0.03197777933544583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,2048,0.00960888879166709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,16384,32,0.009361777868535783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,512,65536,0.02476355599032508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,768,0.004389333228270213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,256,0.003157333367400699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,1024,0.06470488839679293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,5120,0.01974311139848497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,8192,0.06847555769814385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,256,0.01996799972322252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,1024,2048,0.04632711079385546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3584,1024,0.022080888350804646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,3072,0.0075128889746136135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,1536,0.006545777950021956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,65536,0.17276177141401503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,512,0.04853866828812493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,8192,0.2562844488355848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,128,0.005463111317820019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,8192,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,1024,0.005434666656785541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,256,0.008952000074916417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,65536,0.1749502287970649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,51200,0.7836737632751465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,16384,1024,0.010709332923094431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,2560,0.016029333074887592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,256,0.0053093334039052325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,512,0.031176888280444678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,51200,12.791638692220053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2560,6144,0.40131200684441465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,32,0.00853066643079122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3584,768,0.006520888871616787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,51200,0.03499022126197815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,3072,0.0107004443804423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,4096,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,12288,0.04462044437726339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,64,0.008064889245563084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1536,65536,1.020353741115994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,6144,0.052249779303868614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,32,0.003301333429084884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,10240,3.7411155700683594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1536,16384,0.04755999975734287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,5120,0.04302933481004503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,128,0.03625066743956672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,10240,1.1280648973253038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,5120,0.014904000692897372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,5120,65536,0.19360710514916313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,6144,64,0.005979555762476391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,32,0.004505777938498391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,12288,0.020784000555674236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,1536,0.006220444622966979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,7168,0.025019556283950806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,512,0.006182222316662471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,2560,0.30987289216783315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,3584,2560,0.10228800111346775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,51200,512,0.03065511253145006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,5120,0.010487111078368293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,768,0.021617778473430212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,3584,0.05472800135612488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,4096,0.014135999812020196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,7168,0.016183111402723525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,3072,0.1137111054526435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,3072,0.010040889183680216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,64,0.00850222259759903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,1024,0.010943111446168689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,8192,1024,0.06813955307006836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,4096,16384,0.03934222128656175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,10240,0.032080888748168945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,5120,2048,0.07053422265582614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,3072,0.009020444419648912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,128,0.005457777943876054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,7168,0.029600001043743555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,3584,0.01568533314598931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,5120,0.01054222219520145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1536,65536,0.0686826639705234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,10240,0.08746400144365098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,512,0.038344889879226685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,2048,0.006650666809744305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,512,0.005799110978841782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,3584,0.0277839998404185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,768,0.00812266684240765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,2048,0.049581332339180835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,2560,0.03728533453411526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,768,7168,0.21415555477142334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,16384,0.014359111587206522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,1536,0.046675556235843234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,1536,0.8351982434590658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,5120,0.21546222103966606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,256,0.006110222389300664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,768,0.012469333079126147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,6144,0.018616888258192275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,2560,0.08505688773261176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,10240,0.4694231351216634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,128,768,0.013600000076823764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,2048,0.0203804439968533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,10240,0.0795217752456665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,2048,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,32,0.00508266646001074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,6144,0.055236445532904736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,1536,0.00509688879052798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,16384,0.08186488681369357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,12288,16384,0.11044444640477498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3584,5120,0.015319110618697273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,6144,0.19748355282677543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,2560,0.010469333165221743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,32,0.0029173334025674393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,6144,0.1978986660639445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,2560,0.41893956396314835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,64,0.007483555210961237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,1024,0.01627466744846768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,65536,2048,0.7954248852199979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,768,0.005166222237878376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,3072,0.10408088896009658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,512,0.014597333139843412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,16384,0.023794665932655334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,16384,0.4626355701022678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,5120,3584,0.012989333106411828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,8192,0.08948177761501735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,512,0.004552888787455029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,512,0.01015377789735794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,2048,0.0071937781241205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,4096,0.21603200170728895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,6144,0.01926399932967292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,16384,0.04007644454638163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,16384,1.3747644424438477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,3584,0.037725332710478045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,128,0.007485333416197035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,512,0.0574586656358507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,16384,8192,0.05098399851057264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,7168,0.2126000059975518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,6144,0.0063182223174307085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,3072,0.0063528890411059065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,768,768,0.015106666419241162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,768,0.008361777497662438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,12288,0.860786649915907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,12288,0.07519111368391249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,1536,0.010152888794740042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,128,0.005982222242487802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,10240,0.016221332881185744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,32,0.009879111415810054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,16384,1536,0.20941421720716688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,256,0.0047084444926844705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,12288,0.2237928973303901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,8192,0.05363555418120491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,2560,0.014399111270904541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,12288,0.009888889061080085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,3584,0.01053955571519004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,512,0.19729955991109213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,3072,0.022459555003378127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,65536,0.35009333822462296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,512,2560,0.007082666787836287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,768,0.011238222320874533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,2048,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,256,16384,0.011272889044549731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,4096,0.3243315484788683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,6144,0.042860445049073964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,1536,0.011745777395036487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,10240,0.04926400052176582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,12288,0.17672710948520234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,64,0.005261333452330695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2048,16384,0.26389598846435547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,7168,0.3178488943311903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,4096,0.3428035577138265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,5120,0.43087111579047305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,7168,0.0786693361070421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,4096,0.00830666638082928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,768,0.008651555412345463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,65536,7168,0.2367244561513265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,32,0.005283555636803309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,7168,0.06525422467125787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3584,2048,0.012714666624863943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2048,8192,0.20646933714548746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,10240,0.04015644391377767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,51200,25.76242404513889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,2048,0.020032889313167997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,2048,0.03379111157523261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,3584,0.0417537788550059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,256,0.04723822077115377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,16384,1.0414568583170574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,12288,3.779344770643446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,768,0.014729777971903482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,32,2560,0.005660444498062134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,6144,0.013717333475748697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,128,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,51200,0.08909333414501613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,5120,6144,0.10372177759806316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,5120,2048,0.012460444536474017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,4096,0.10321244266298081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,2048,0.009667555491129557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,1536,0.0435804459783766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,2560,0.013487999637921652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3584,6144,0.15652533372243246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,16384,0.03644977675543891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,5120,0.02108888824780782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,2048,0.05407555566893684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,7168,0.03944089015324911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,10240,0.016403555870056152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,1536,0.006511111226346757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,768,2560,0.019866666860050626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,32,0.002699555622206794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,8192,0.5350435574849447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,3584,0.0656906697485182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,6144,0.08407021893395318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,10240,0.12617599964141846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,51200,0.07080177466074626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,6144,0.010625777973069085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,12288,0.035178667969173856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1536,8192,0.01402844488620758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,5120,0.009672889278994666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,2560,0.00637511122557852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,51200,0.008260444634490544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,32,8192,0.011896000140243106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,2560,0.007320000065697565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,7168,0.025821333130200703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,512,0.019845333364274766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,65536,0.009528888596428765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,2560,0.006227555374304454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,12288,0.019063111808564927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,2048,0.006659555352396435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,4096,0.048093332184685596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,51200,0.018560000591807894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,12288,0.017705778280893963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,2048,0.08003199762768216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,3072,0.6358639929029677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,65536,32,0.24655289120144316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,12288,6144,0.3168746630350749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,768,0.009633777870072259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,51200,5120,3.4388870663113065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,10240,0.017775999175177682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,7168,0.0184560004207823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,6144,0.02852355440457662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,8192,0.09416000048319499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,512,0.006322666588756774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,3584,0.04791733291414049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,1024,0.004132444245947732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3072,8192,0.015399111641777886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,768,65536,0.3055555555555555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,2560,0.042575998438729175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,256,0.09985333018832737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,8192,0.19768533441755506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,64,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,6144,0.013136000268989138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,65536,0.25892265637715656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,1536,0.056553781032562256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,16384,0.026334222820070054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,65536,0.009097778134875828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,4096,0.18983733654022217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,2048,0.02589955594804552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,2560,0.035790221558676824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,768,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,6144,0.23654132419162324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,768,0.007328000333574083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,1536,0.14016800456576878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,65536,0.15510666370391846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,64,0.005946666830115848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,10240,0.34922311041090226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,32,0.0030533334033356774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,256,0.15764089425404867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,3584,0.19600178135765922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,16384,0.0274008893304401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,32,0.002667555585503578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,12288,0.2172328895992703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,5120,0.010839111275143094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,2560,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,5120,0.46347464455498594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,2560,0.023367111881573994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,4096,0.013771555489963956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,51200,2048,0.07362044519848294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,51200,0.3432151211632623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,1024,0.015452444553375244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,12288,0.05305066704750061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,7168,0.018274666534529794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,1024,0.006211555666393704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,3584,0.014494222071435718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,16384,0.8209022416008843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,51200,0.1517431073718601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,256,0.15278132756551108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,64,0.032906668053732976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,8192,0.05039733317163256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,16384,0.13729777601030138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,512,0.004391111019584867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,2560,0.013615999784734515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,5120,0.03400177756945292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,51200,0.07014133532842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,3584,0.01868799991077847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,6144,3072,0.01975911193423801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,12288,0.558858659532335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,128,768,0.005104889058404499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,51200,0.05899644560284085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,6144,0.011523555550310345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,256,2560,0.01203733351495531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,128,0.005764444255166584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,3584,0.008462222086058723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,256,0.011343110766675739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,1536,0.051853332254621715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,3584,0.05932088692982992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,51200,0.04797333478927612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1536,51200,0.4626808696322971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,51200,0.05378489030732048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,7168,0.3137626647949219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,2048,0.007111111448870764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,51200,0.06417955292595758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,12288,0.057908442285325795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,1024,0.012619555824332766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,10240,65536,0.5595413313971626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,128,768,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,1536,0.008009778128729926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,2048,0.009172444542249044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1536,8192,0.1448062260945638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,3072,0.009359999663299983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,8192,0.10503911309772068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,3584,0.03074044320318434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,3584,0.014942222171359591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,1024,0.0057751110030545135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,4096,12288,0.03127022253142463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,1536,0.10467911428875393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,65536,1.0712862014770508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,5120,0.015238222148683337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,768,3072,0.0414684447977278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,128,0.0037875556283526947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,8192,51200,0.5401235686408149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,51200,0.20835467179616293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,128,0.009863999982674917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,5120,0.010084444450007545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,1536,0.05223555697335137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,6144,0.03173689047495524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,128,12288,0.008210666477680206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3584,8192,0.08105066749784681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,7168,0.1525884469350179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,2048,0.010423111418883005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,32,0.0026897777699761917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,6144,0.036547554863823786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,10240,0.06733510891596477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,768,256,0.0058515556156635284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,51200,0.1497946712705824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,512,0.6192746692233616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,2560,0.007157333195209503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,2048,0.0075937774446275495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,3584,0.010312000082598792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,51200,0.27162665790981716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,12288,0.039721777041753135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,12288,51200,5.095545874701606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,768,0.006119110931952794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,51200,0.06390488810009427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,1536,0.10089688830905491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,128,0.003152888889114062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,4096,0.011178666518794166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,5120,0.4618364440070258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,65536,0.9084097544352213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,4096,0.008626666333940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,10240,0.05770666731728447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3072,4096,0.01371911085314221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3072,3584,0.08593333429760402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,8192,512,0.01900800069173177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,10240,0.035290665096706815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,51200,3072,0.05836621920267741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,10240,0.138319108221266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,6144,0.07082755698098077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,8192,0.1294648912217882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,2048,0.01572088897228241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,2048,0.040228443013297185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,128,0.006799110935793982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,51200,0.27195821868048775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,128,0.00924355536699295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,64,0.006685333533419504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,3072,32,0.002584888910253843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3072,6144,0.6800933414035373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,4096,0.010989333192507425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,32,6144,0.01846311158604092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,51200,65536,1.7825751834445531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,51200,4.4449441697862415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,32,0.0032071111102898917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,3072,0.011501332951916588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,768,0.04122844338417053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,16384,0.021709332863489788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,65536,0.012173333101802401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,768,0.0057857777509424425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,1024,0.008252444366614023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,3584,0.13700444168514675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,128,0.03587822119394938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,65536,0.009735999835862054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,12288,0.050028443336486816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,12288,6.043287913004558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,3072,0.03377244538731045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,7168,0.7170640097724067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,51200,64,0.006626666833957036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,3072,0.19777066177792021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,51200,0.3817217879825168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,2048,0.06019022067387899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,51200,0.19407378302680123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,1024,0.005967999911970562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,4096,1.4785955217149522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,1536,0.00434933313065105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,128,0.00556355549229516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,2048,0.021048888564109802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,16384,6144,0.06685511271158855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,5120,4096,0.6475893126593696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,8192,0.025770665870772466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,6144,0.018725333942307364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,32,0.005946666830115848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,10240,0.014447111222479077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,12288,0.15572978390587702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,8192,0.01204711116022534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,10240,0.0923137797249688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,5120,0.014869333969222175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,3072,0.1562017732196384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,1536,0.02674399978584713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,32,0.005330666485759947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,7168,512,0.03760711020893521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,4096,0.009903111391597325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,2048,0.011905777785513135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,8192,2.63047112358941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,256,0.00674044465025266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,6144,0.14835822582244873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,16384,5120,0.354056888156467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,768,0.004011555678314633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,5120,0.023759111762046814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,4096,0.034416000048319496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,768,0.03452355662981669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,65536,0.04407733347680834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1536,7168,0.09607644213570489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,10240,2.763887193467882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,64,7168,0.0141004439857271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,6144,0.5179493162367079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,64,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3584,65536,0.0854826635784573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,8192,0.09043377637863159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,768,0.3234817716810438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,2048,0.1880648930867513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,5120,64,0.011691555380821228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,12288,0.053322666221194796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,8192,0.12114132775200738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,1024,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,768,0.01350400017367469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,2560,0.005945777727497949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,4096,0.07373866770002577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,4096,0.015853333804342482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,12288,0.09372978077994452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,16384,1.53711732228597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,768,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,16384,0.026363554928037856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,16384,16384,1.240912013583713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,16384,0.2329111099243164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,8192,0.015112888481881885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,1536,0.11280622747209336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,65536,0.11317333910200332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,12288,64,0.0031146665828095544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,256,0.004040889028045866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,3584,0.029853334029515583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,51200,0.40530843204922146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,7168,0.23603910870022246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,10240,7168,0.31559912363688153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2048,8192,0.025040000677108765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,4096,0.019908444748984445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,128,0.024449777272012498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,2048,0.016536888149049547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,64,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,7168,0.1305217742919922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,32,0.10384355651007758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,256,0.012504888905419244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,512,0.005412444472312927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,768,0.013248888982666863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,12288,0.07215022378497653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,256,0.09025600221421982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,6144,0.010417777631017896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,4096,0.05138488941722446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,3072,0.006310222049554189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,8192,0.01310400002532535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,8192,0.01148088855875863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,51200,3.239701377020942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,7168,0.08099289072884454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,6144,0.015231110983424716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,4096,3584,0.013331555657916598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,768,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,32,0.005150222116046482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,512,0.003988444391224119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,2560,1024,0.01812977757718828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,16384,1.1782044304741752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,512,0.353091557820638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,32,0.0033768889390759995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,128,0.006633777585294511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,12288,0.21741955810123018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,4096,0.6885840098063151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,65536,2048,0.7435830963982476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,4096,0.013872000078360239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,128,0.006978666616810693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,16384,0.010076444182131026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,1024,0.1285288863711887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,51200,0.09820444716347589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,1024,0.004911111046870549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,3584,0.014463111758232117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,7168,0.24575556649102104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,128,0.0030533334033356774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,12288,65536,2.8201323615180125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,3072,0.008723555339707268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,256,0.006679111056857639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,16384,0.02537600033813053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,32,0.003257777748836411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,2560,0.04912977748446994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,3072,0.023756444454193115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2560,256,0.012368000215954252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3072,2048,0.033542222446865506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,3584,0.024119999673631456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,1536,0.00923999978436364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,7168,0.01698933376206292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,6144,0.04124888777732849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,1024,0.006370666540331311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,1536,0.006029333505365584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,51200,0.0635511080423991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,8192,0.021984888447655573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,51200,4.086513943142362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,3584,0.008606221940782335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,7168,0.035980443159739174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,12288,0.3343848917219374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,5120,0.04919644527965122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,256,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,2048,0.027296000056796606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,2560,0.018144889010323417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,2048,0.023440000083711412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,5120,1.1617502636379666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,256,0.006247111078765657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,64,0.003273777870668305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,3584,0.008927110996511247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,51200,2.3876274956597223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,65536,6.462156507703994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,51200,0.9832017686631945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,3584,0.021260443660948012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,10240,2.1234363979763455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,5120,64,0.0030951110853089225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2048,5120,0.011407111254003314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,3584,0.01017777787314521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,8192,0.06132355663511488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,5120,0.027686221732033625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,16384,0.3862568802303738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,128,0.008204444415039485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,5120,0.016047999262809753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,512,0.008338666624493068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,16384,0.8369386461046008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,32,0.0515342222319709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,65536,0.22566755612691244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,6144,0.01881866653760274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,1536,0.0064151109092765385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,32,0.00608800020482805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,768,10240,0.09771821896235149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,8192,8192,0.1972524457507663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,512,0.005366222312053044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,7168,0.23156711790296766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,4096,0.008552889029184977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,5120,0.7181759940253364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,128,0.024689777029885188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,3584,0.020964443683624268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,51200,3.520426644219292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,65536,51200,1.7534639570448134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,8192,0.010096888575288985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,51200,0.17954488595326742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,3072,0.0322035551071167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,51200,0.4021635585361057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,512,7168,0.01093155559566286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,512,0.01794666714138455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,16384,0.026783999469545152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,1024,0.18081333902147081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,256,0.015670221712854173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,128,0.003014222201373842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,5120,0.052067554659313627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,2048,0.026319111386934917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,3584,0.08602666854858398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,12288,0.039353777964909874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,256,5120,0.006704000135262807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,64,0.05084355672200521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,3072,32,0.011367999845080905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,256,3584,0.007246221933099959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,4096,0.12164889441596137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,1536,0.04256177941958109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,65536,0.23140533765157065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,12288,0.020430222153663635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,1536,0.028775112496482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,4096,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,65536,1.0625386767917209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,3072,0.9304328494601779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,512,0.004831111265553368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,10240,0.9905867046780057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3584,12288,0.18696088261074492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,64,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,1024,0.0041786668201287585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,2560,0.012846222354306115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,512,0.019090667366981506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,128,0.01461955573823717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,128,0.006223111102978389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,2560,128,0.004878222114510006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,4096,0.03878577881389194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,512,0.004867555366622077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,512,0.01980533368057675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,256,0.03574577636188931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,65536,0.7340853479173449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,3584,0.033737778663635254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,16384,256,0.005283555636803309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,16384,0.7732791370815701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,768,0.003924444317817688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,128,0.003134222287270758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,3072,0.3096071084340413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,768,16384,0.021519111262427434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,6144,0.1423919995625814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,512,0.0047662220895290375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,10240,0.010265777508417765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,7168,2.3582710689968533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,6144,0.3700711197323269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3584,64,0.010070222119490305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,51200,0.06978044244978163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,3072,2560,0.05551377932230631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,4096,0.11510666211446126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,7168,5120,0.017080888152122498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,7168,0.02040799955526988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,7168,3584,0.09898577796088324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,3072,0.18998400370279947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,65536,5.012402852376302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,16384,0.049719111786948315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,768,0.05106133222579956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,768,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,7168,0.012099555797047086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,16384,2.5287679036458335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,768,0.02370400064521366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,2560,0.022886221607526142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,32,0.011583999627166323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,3584,0.009840000006887648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,7168,3584,0.1284773349761963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,51200,8.94923824734158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,4096,0.024303999212053087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,8192,6144,0.02683555583159129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,6144,0.23180890083312988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,32,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,768,0.005764444255166584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,1024,0.023040889037979975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,2560,0.12946755356258816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,51200,0.061956445376078285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,3584,0.1905368831422594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,10240,0.06889777713351779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,5120,0.5233564376831055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,5120,0.030574222405751545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,10240,0.5511537657843696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,6144,0.6687217818366157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,16384,0.10389777686860825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,512,0.00611111107799742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,4096,0.007099555598364935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,768,0.015113777584499784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,1024,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,6144,0.011547555526097616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,128,0.01759377784199185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,16384,0.06169422467549642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,7168,0.011482666763994427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,51200,51200,0.7155289120144315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,51200,7.681321038140191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,64,65536,0.06678577926423815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,2560,0.04138133260938857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,6144,0.4085991117689345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,12288,0.39883377816942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,5120,0.17035467094845244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,8192,0.2460399998558892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,7168,0.011562666959232755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,12288,0.8946373197767469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1024,7168,0.012991999586423239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,10240,0.020463999774720933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,16384,0.06312621964348687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,256,0.00342755557762252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,10240,0.7309528986612955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,3584,0.009268444445398118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,3072,0.00847999999920527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,512,0.008433777425024245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,64,0.004357333398527569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,3072,0.010630221830474006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,6144,7168,0.20366311073303223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,1024,0.00889155517021815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,65536,0.11255021890004475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,64,768,0.003863999827040566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,10240,0.033599111768934466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,5120,0.020083554916911654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,65536,0.07054044140709771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,64,0.004671111288997862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,3072,0.05451822280883789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,32,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,10240,0.048431111706627734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,256,51200,0.4917439884609646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,16384,0.25387199719746906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,128,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1024,8192,0.009837333526876237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,4096,0.32391201125250924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,65536,0.24415198961893717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,8192,0.016303999556435477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,16384,0.03358488943841722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,4096,0.3099964459737142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,64,0.007490666376219855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,2048,0.011007111105653973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,3072,0.013183111117945777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2048,1536,0.03914666506979201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,10240,0.031376888354619346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,32,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,128,3072,0.006709333509206772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,7168,512,0.376418670018514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,1536,0.021244444780879553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,768,0.006533333410819371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,2048,0.02834933333926731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,768,0.007245333658324347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2560,2560,0.008361777497662438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,3584,0.009036444127559662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,3072,0.029217776325013902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,6144,0.010965333216720156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,6144,256,0.0059164443777667145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,2048,1.2491209242078993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,256,0.004888888862397936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2560,6144,0.019304000669055514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,6144,0.0336773329310947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,768,0.007814221911960179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,64,0.009512888888518015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,3584,0.04608266552289327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3584,2048,0.013033777475357056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,3584,0.007896000312434303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,3072,0.06308355596330431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,51200,18.797314114040798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,4096,0.015807999504937064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,768,0.004875555634498596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,8192,0.05738933218850029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,128,0.0028284444577164124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,7168,0.5054293208652073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,12288,0.6558177736070421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,4096,0.011704000333944956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,4096,0.006974221931563483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,7168,0.11322132746378581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,51200,0.7848018010457357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,512,3.1152222951253257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,64,0.00332266671790017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,2560,0.02202577723397149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,4096,0.212482664320204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,2048,0.01276088919904497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,512,2560,0.00685777763525645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,64,0.0028328889360030494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,6144,0.3220737775166829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,51200,0.27808888753255206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,5120,1.2821822696261935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,512,0.006794666664467917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,3072,0.007610666255156199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,512,0.029616889026429918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,4096,0.007114667031500075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,6144,0.02919644448492262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,128,0.036943111154768206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,128,0.003321777822242843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,65536,0.05400800042682224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,1536,0.011679111255539788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,10240,0.1741351154115465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,2560,0.015607110328144498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,256,0.13254577583736843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,256,0.016327111257447135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,65536,0.29569776852925617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,5120,0.055550220939848155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,3072,0.007373332977294922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,64,10240,0.03772799836264716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,10240,0.01256533298227522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,1024,0.012988444831636218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,5120,0.04516622092988756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,7168,0.7534755600823296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,768,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,1024,0.01203555530971951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,10240,1.2181066936916776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3584,512,0.004146666576464971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,51200,0.11697244644165039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,512,0.03991466760635376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,3072,0.033878223763571844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,2560,0.00850666645500395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,8192,0.028413332170910303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,512,0.055606219503614635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,65536,0.08928889036178589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,128,0.0053546664615472155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,16384,0.2507644494374593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,65536,1.9425040351019964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,512,0.012358222570684222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,51200,0.0605564448568556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,768,0.022970666488011677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1024,256,0.00628977765639623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,2048,0.010261333651012844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,2048,0.011767999993430244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,51200,1.5840319527520075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,10240,0.08257066541247897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,12288,0.40014134513007266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,65536,1.4676897260877821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,2048,0.007692444655630324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,5120,0.07453244262271456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,1536,0.03194844391610887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,3072,0.01165422217713462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,256,0.007838221887747446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,3072,256,0.005399999933110342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,2048,0.055488000313440956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3072,12288,0.3247751129998101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,2048,0.041952000723944775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,1024,0.007606222397751278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,16384,0.23717776934305826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,51200,16384,0.2438417805565728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,10240,0.025711110896534387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,3072,0.013321778012646569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,8192,128,0.05966933568318685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,8192,768,0.0841919978459676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,2560,0.05268355541759067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,10240,5120,0.13363822301228842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,51200,3072,0.5098346604241265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,16384,0.025773333178626165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,3072,0.008425777984990014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,2048,0.0069360000391801195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,7168,0.058279110325707324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,4096,5120,0.015690666106012132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,12288,0.4164293342166477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,2560,0.01216533366176817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,6144,0.010517333116796283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,3584,0.02508977717823452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,10240,51200,1.7830551995171442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,64,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2560,3584,0.009375111096435124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,768,0.12482310665978326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,8192,0.2284017668830024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,64,0.01703555550840166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,256,128,0.003879111260175705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,1536,0.0432275566789839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,3072,0.02474044428931342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1024,6144,0.04459733433193631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,5120,0.013919111755159167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,51200,0.08431022034751044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,1536,0.061271111170450844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,65536,0.06557244724697538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,8192,64,0.006712889091836081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3072,32,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,512,0.004601777841647466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,512,512,0.004501333253251182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3584,512,0.018999111321237352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,16384,0.24090043703715006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,512,0.011662222445011139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,6144,0.18681955337524414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,512,0.0057813334796163775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,768,0.04673777686225044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,16384,0.021998221675554912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,2048,7168,0.018064889642927382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,4096,128,0.004716444346639845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,8192,0.009117333425415887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,3584,0.016513778103722464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,5120,0.18389777342478433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,6144,3584,0.21515022383795845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,3072,0.008622222476535374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,8192,0.01996799972322252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,64,1024,0.004072000169091755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,2560,0.00739911115831799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,1024,0.02561155623859829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,2560,0.007827555967701806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,768,0.003474666840500302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,8192,0.03507644600338406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,2048,0.010856000085671743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,128,0.005431111074156231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,128,32,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1024,512,0.004443555656406614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,1536,0.05407377746370104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,4096,0.10725688934326172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,768,0.0059262220230367445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,12288,0.021988444858127173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2048,3072,0.007660444411966536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,51200,0.14523200194040933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,6144,0.08255377742979261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,128,0.003866666720973121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,1536,0.004916444420814514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,3072,0.049623999330732554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,2560,0.007677333222495184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,51200,0.21270222134060332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,10240,0.05967555443445841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,128,0.003069333318206999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,32,0.18175466855367026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,32,0.05566666523615519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,7168,0.1316186719470554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,128,5120,0.00638488887084855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,10240,1536,0.12354132864210342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,128,0.013040888640615674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,8192,0.019900444481107924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,7168,0.20305155383216011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,4096,0.05158311128616333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,768,0.008833777573373582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,8192,768,0.009240888886981541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,51200,5120,2.2349537743462458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,5120,0.0129457778400845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,64,0.005466666486528184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,64,0.005653333332803514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,7168,0.014064888159434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,12288,0.13872178395589194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,5120,8.988904317220053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,128,256,0.010024888647927178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,65536,0.37532001071506077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,1536,0.010057777994208867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,512,0.07302221986982557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,2048,1024,0.004439110971159405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,768,0.039790223042170204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2560,51200,0.07709778017467923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,3072,0.02754755483733283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,8192,1.2792978286743164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,65536,5.012967003716363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,8192,0.03496888942188687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,256,0.1257057719760471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,10240,0.08051466941833496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,128,65536,0.093322667810652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,7168,7168,0.020958221620983545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,32,0.012030222349696688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,2048,0.006068444500366847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,5120,0.031799111101362444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,32,4096,0.006031111296680238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,1536,0.005312888986534542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,512,0.12484710746341282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1536,5120,0.04042311178313361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,12288,0.2083973354763455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,32,5120,0.009544889132181803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,512,0.010083555347389644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,3072,0.035346666971842446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,3072,0.02884355518552992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,6144,0.16671021779378256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,768,0.005532444351249271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,12288,512,0.053704000181621976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,1024,0.005712000032265981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,1024,0.006504000061088138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,2560,0.02922755479812622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,12288,0.6485697958204482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,7168,0.013296888934241401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,7168,0.3156489001380073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,3072,0.056484447585211865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,1536,0.006868444383144379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,10240,0.3308933311038547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,6144,0.7538702223036023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,65536,0.09104177686903213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,7168,0.018343110879262287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,4096,12288,0.03254488772816128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,51200,27.664072672526043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,10240,1.9516133202446833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,51200,3.5960053337944875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,512,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,7168,0.02662044432428148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,256,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,10240,8192,0.22429866260952422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,256,0.10708177751964992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,5120,0.21258666780259874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,3584,0.018089777893490262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,32,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,4096,12288,0.04123555620511373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,1024,0.019996444384257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,512,0.0037342223028341928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,3584,0.02086044516828325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,128,0.005048888838953442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,512,0.005422222117582957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,8192,0.14480177561442056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,65536,3072,0.06238310866885715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,51200,0.14271999730004206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,1536,0.0046942221621672315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,512,1.1237422095404732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,32,0.013972444666756524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,51200,3584,2.2749911414252386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1536,51200,0.06882755623923408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,10240,1.0583261913723416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,2560,0.008088889221350351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,1536,0.007821333077218797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,12288,0.020584000481499564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,1536,0.06439644760555692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,2560,0.017995556195576985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,1024,0.006021333237489064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,64,0.005691555639108022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,3072,0.010499555203649731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3072,1024,0.01908533275127411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3584,3072,0.08008178075154622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,3584,0.015602666470739575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,32,0.015648888217078317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,1536,0.006425777657164468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,3584,0.07193422317504883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,512,0.005398222141795688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,768,0.004997333304749595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,1024,0.005490666462315454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,768,0.01757066614098019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,3072,0.012882666455374824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,16384,0.05217599868774414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,5120,0.009673777553770278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,256,0.005044444567627377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,5120,0.028629332780838013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,1536,0.00720266666677263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,1536,0.007378666765160031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,128,0.00434400017062823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,2048,0.05944710969924927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,2560,0.3119155565897624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,65536,23.412801106770832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,16384,0.10790310965643989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,6144,0.020626667473051283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,3072,0.00962933318482505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,2048,0.015504888362354703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,256,8192,0.03689333465364244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,1536,0.014749333262443542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,12288,512,0.01037599974208408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,512,256,0.004359111189842224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,2048,0.011028444601429833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1024,8192,0.7658461994594998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,12288,0.18775910801357695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,768,256,0.004780444420046276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,32,0.003738666574160258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,3584,0.04710400104522705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,3072,0.04139466749297248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,64,10240,0.007445333732499017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,2048,0.01555377741654714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,32,0.0027013334135214486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,12288,0.08661688698662652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,128,2560,0.006004444426960415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,3072,0.00868088917599784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,1536,0.06744177473915948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,256,16384,0.019621334142155118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,256,768,0.005159111072619756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,256,0.015609777635998197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,16384,0.03147111005253262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,768,256,0.005690666536490123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,128,0.005229333208666907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,6144,0.010119111173682744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,512,0.04828177889188131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,16384,512,0.10686755842632717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,256,0.15234933959113225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1024,3072,0.011169777976142036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,65536,0.0741564432779948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,2048,0.04509777824083964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,1024,0.005470222069157495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,1536,0.009848888549539778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,2560,0.013905778527259827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,256,0.06194577614466349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,256,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,16384,0.03778310947948032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,6144,0.02011644509103563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,32,0.004369777937730153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,5120,0.025640888346566096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,51200,5120,0.11969867017534043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,3584,1.5960755878024633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,256,0.006836444553401735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,12288,0.04287199841605293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,2048,0.7897333568996854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,2560,0.10155289040671454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,512,0.004763555609517627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,64,0.009697777529557547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,12288,0.026826666461096868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,51200,10240,2.83461062113444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,5120,0.018175111876593698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,256,0.0042462220622433555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2048,1536,0.008573333422342936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,1536,0.00664444433318244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,10240,0.025057777762413025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,1024,0.013580444786283704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,512,0.01647111111217075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,6144,0.013658666776286231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,7168,0.014931556251313953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,4096,10240,0.027136888768937852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,4096,1024,0.009167110754383935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2560,2560,0.03532977898915609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,256,0.0037164443896876443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,10240,0.12985332806905112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,4096,0.11445600456661648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,16384,12288,0.4516124195522732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,51200,1.4810124503241644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,2560,0.007265778051482306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,3072,0.01143200033240848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,10240,1.9946515825059679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,1024,16384,0.029184887806574505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,4096,1024,0.008538666698667739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,32,512,0.004322666674852371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,6144,0.1710124413172404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,6144,0.3027217653062609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,6144,0.013443555268976422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,7168,0.2741182115342882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,6144,0.008759111166000366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,1536,0.1118275589413113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,2048,0.006748444504208035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,256,0.003089777711364958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,768,4.381880866156684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,128,0.003547555456558863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,8192,0.14166133933597141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,4096,0.01870577699608273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,2560,0.022727110319667395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,8192,0.5492329067654079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,2560,0.02858577834235297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,4096,0.011151110960377587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,65536,0.6069964302910699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,768,256,0.004931555440028508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,64,51200,0.05964622232649061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,512,0.00608711110221015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,16384,5120,0.026965333355797663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,2560,0.013063111239009433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,1024,0.009892444643709395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,7168,0.290154669019911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,2048,0.008058666355080074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,128,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,256,0.003119999956753519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,51200,64,0.10805155833562215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,3072,0.0402764446205563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,51200,512,0.015768888923856948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,2560,0.006237333433495627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,7168,0.26097332106696236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,128,65536,0.8136773109436035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,512,0.01422133379512363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,7168,0.058874666690826416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,5120,0.016343111793200176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,32,0.0025946667624844448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,64,0.0029191111938820947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,5120,0.04427999920315213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,5120,8192,0.031587554348839655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,32,0.002518222149875429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,2048,0.0053671110007498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,4096,3584,0.12353777885437012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,2560,0.008607111043400234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,2048,0.03666311171319749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,64,0.005316444569163852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,5120,0.040399110979504056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,128,2560,0.019514666663275827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,51200,0.031397332747777305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,2048,0.015589333242840238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,3584,0.013580444786283704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,8192,8192,0.11627999941507976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,2048,0.009158222211731805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,16384,0.04989155464702182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,4096,0.007882666256692674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,7168,0.05437511205673218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,768,0.005593777944644292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,64,0.003371555565132035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,8192,0.2119599978129069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,5120,0.02090577781200409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,5120,51200,1.8568905724419489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,8192,0.09762577878104316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,3584,0.04044977823893229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,32,0.0026515554636716843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,7168,12288,0.03083555565940009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,256,6144,0.025448888540267944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,32,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,10240,0.01888177792231242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,2560,0.011041777829329172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,32,0.0026026666164398193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,512,0.0033777778347333274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,32,128,0.00272444449365139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,5120,0.01089422239197625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,6144,6144,0.24108534389071992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,7168,0.017136888371573556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,256,0.005710222240951326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,512,0.03613155417972141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,12288,0.019552888141738046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,4096,15.041715833875868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,2560,0.06728088855743408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,2048,0.010392888552612728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,8192,0.022462222311231825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,51200,0.062048892180124916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,4096,0.016889777448442247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,1536,0.010571555958853828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,5120,0.02665333284272088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,768,0.0057795556883017225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,32,0.0030533334033356774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,768,0.005537777725193236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,2048,0.00943999985853831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,8192,0.13410666253831652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,256,0.033032887511783175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,2048,0.007131555842028723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,1024,0.010302222437328762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,2560,0.007003555695215861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,768,65536,0.6262577904595269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,12288,64,0.0033119999700122406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,128,0.0075048887067370946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,12288,0.08709600236680771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,8192,0.13296711444854736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,5120,2560,0.06290133131874932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,8192,0.4555857976277669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,10240,0.03051466743151347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,32,0.0030951110853089225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,512,0.0063982220987478895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,64,0.007294221884674496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,768,0.0047253333032131195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,64,0.025088000628683303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,6144,0.011366222467687396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,8192,0.019334221879641216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,12288,32,0.01555022266176012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,6144,0.08488000101513332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,1536,0.005519999812046687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,256,0.003801777958869934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,32,0.007061333292060428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,256,0.005769777629110549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,8192,65536,3.379088931613498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,12288,0.09599110815260146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,8192,12288,0.04054844379425049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,16384,0.07653866873847114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3584,3584,0.07310577895906237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,6144,0.17421244250403511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,12288,64,0.007517333659860823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,16384,0.018333334061834548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,8192,0.04785511228773329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,256,512,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,3072,0.009939555492666032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,10240,0.0214720004134708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,32,0.0027884443600972495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,7168,0.30526667171054417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,64,0.0027351110345787476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,768,0.01885777711868286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,768,0.008414222134484185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,65536,0.8944248623318143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,12288,0.028981334633297388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,6144,0.01183644433816274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,2048,0.0759048859278361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,10240,3584,0.01943555474281311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,32,0.006495999793211619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,65536,2048,0.516341315375434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,3584,0.00924622184700436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,128,0.0074986666440963745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,10240,0.18239111370510527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,128,0.004649777793222004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,512,6144,0.012376888758606382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,6144,0.027031110392676458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,768,0.006004444426960415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,3584,0.010591111249393888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,128,2048,0.005280888742870755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,10240,0.030076444149017334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,8192,0.02037244372897678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,6144,0.08807110786437988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,2048,3072,0.01035288886891471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,5120,0.008897778060701158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,768,0.05154755711555481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,12288,0.015023999743991427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,512,256,0.004591111093759537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,1024,0.08603644371032715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,51200,0.1864888932969835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,256,4096,0.008299555215570662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,128,8192,0.02460799945725335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,6144,0.03655644589000278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,768,0.016898666818936665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3584,3584,0.07946044206619263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2560,256,0.004892444445027245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,128,5120,0.00962222201956643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,7168,0.048671109808815845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,32,0.002597333242495855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,10240,0.04012444284227159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,512,32,0.0041475556790828705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,2560,0.06782311201095581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,128,0.011693333586057028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,1024,0.01131377783086565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,2560,6144,0.017589333984586928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2048,2560,0.007140444384680853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,1536,0.029525332980685767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,2048,0.5461511082119411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,4096,0.033357332150141396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,51200,0.05966933568318685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,1024,5120,0.028910222980711196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,51200,5.2841572231716585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,512,256,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,65536,5.30072996351454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,768,0.006959999601046245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,51200,0.25347556008232963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,768,0.004384888956944148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,3584,0.007361777954631382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,1024,0.012847111456924014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,64,0.07876977655622694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,128,0.016880000631014507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,1536,0.2013253370920817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,128,0.01868533260292477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,8192,0.39375911818610293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,3584,0.011609777808189392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,5120,0.22673334015740287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,6144,1536,0.012667555775907306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,512,0.006231110956933763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,16384,5120,0.057408001687791615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,16384,0.15795111656188965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3072,51200,0.5129342079162598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,3072,0.12085066901312934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,1024,0.01761244402991401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3072,2560,0.008700444466537898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,5120,0.019108444452285767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,32,12288,0.04618044363127815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,512,0.005615111026499007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,4096,0.657491577996148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,6144,0.03133066826396518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,64,0.0026524443593290118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,2048,0.006861333217885759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,6144,0.029673778348498877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,10240,0.31389954355027944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,512,768,0.0038604442444112566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3584,3072,0.010776888993051318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,768,0.044453332821528115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,4096,0.10918222533331977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,12288,0.015523556205961438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,128,0.04704977737532722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,65536,0.011393778026103973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,12288,2.563945770263672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,1536,0.18348444832695854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,4096,1.0988684760199654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,5120,0.036079999473359846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,1536,0.01815555493036906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3584,128,0.025571554899215698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,65536,0.5511520173814561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,128,0.0058373332851462895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,5120,0.009843555589516958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,65536,0.10992533630794948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,16384,5120,0.04646133383115133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,7168,0.2067955599890815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,8192,0.07041511270734999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2048,1536,0.012308444413873883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,768,1536,0.01791466606987847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,65536,0.2686071130964491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,512,768,0.015923556354310777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,16384,0.031164444155163232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,2048,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,65536,0.7176311280992297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,5120,0.16980444060431588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,4096,0.01740977830357022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,8192,0.1048568884531657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,2560,2560,0.008930666579140557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,4096,0.2699226538340251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2048,16384,0.027024888330035742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,2048,0.04932622114817301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,64,0.00620177760720253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,51200,0.1115564505259196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,10240,0.06528444422615899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,65536,1536,0.28420623143513996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,65536,0.14501778284708658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,768,16384,0.09231022331449722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,7168,0.7745653258429633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,4096,0.018361777067184448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,12288,0.060067554314931236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,2048,0.03384800089730157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1024,2560,0.006802666518423293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,12288,0.4494640032450358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1024,256,0.004659555438492033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,3072,0.7002888785468208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,65536,512,0.4065511226654053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,12288,3584,0.025618667403856914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,16384,0.08931377861234878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,8192,0.10359999868604873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,4096,0.008135110967689091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,128,0.0075528886583116316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,10240,1024,0.06840622425079346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,512,0.003535110917356279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,1024,0.009742221898502773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,7168,0.0330079992612203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,512,1536,0.006401777681377199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,65536,0.3845955530802409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,3584,0.02216622233390808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,12288,4096,0.020427554845809937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,1024,0.05815200010935465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2560,16384,0.05967555443445841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,512,0.004886222382386525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,4096,0.020398222737842135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,10240,0.30485155847337514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,1536,0.10872711075676812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,6144,0.007812444534566667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3584,3584,0.07497866948445638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,2560,0.0070364442136552595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,3584,0.010750222537252637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,32,64,0.00646577775478363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,8192,0.12406044536166722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,512,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,32,0.0027760000278552375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,6144,0.13158399528927273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,1536,0.012912888493802814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,6144,0.018980445133315194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,128,0.003638222399685118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,512,0.018952889574898612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,16384,0.082241780228085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,65536,0.04132711225085788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,1024,0.00756444450881746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,16384,0.6571120156182183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,256,0.01388711151149538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,10240,64,0.007334222396214803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,768,0.005723555468850666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,256,0.0034799998005231223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,1536,0.008218666745556725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,512,0.21920622719658744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,64,0.0027919999427265594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,5120,256,0.03224710954560174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,51200,0.13546578089396158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,51200,0.08455555306540595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,8192,0.015040889382362366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,2048,0.019279999865425956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,768,0.005056000004212062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3584,4096,0.010703110860453712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,3584,0.01644977761639489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,12288,0.40259112252129453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,2560,0.007073777417341868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,2560,0.010087110930018956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,3584,0.009249777429633671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,1536,0.0752097765604655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,65536,0.7465884420606824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,7168,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,8192,0.03396000133620368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,2560,0.3730213377210829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,65536,0.29827822579277885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,6144,0.010478222535716163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,32,0.005538666827811136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,512,0.008416888614495596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,256,0.00925511121749878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,8192,0.2197128931681315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,10240,0.14227733347151014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,64,0.1053768859969245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,3072,0.015288000305493673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,7168,256,0.023637332850032385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,65536,0.10660355620914036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,16384,2.4886834886338973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,51200,256,0.6158320109049479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,6144,0.20509866873423258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,10240,2048,0.014515555567211576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,5120,0.010443555812040964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,65536,1.8214035034179688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,7168,0.03109955456521776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,128,0.022767111659049988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,2560,0.007496889266702864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,5120,0.011949333051840464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,512,0.016136889656384785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,3072,0.028328001499176025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,64,0.0188071115149392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,7168,0.1868293285369873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,1536,0.00674222202764617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,51200,12288,0.7683768802218967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,3584,0.3185928927527534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,128,0.00508266646001074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2560,3072,0.059205333391825356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,3072,0.1689137750201755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,65536,128,0.007500444849332173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,4096,0.14118844932980007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,128,0.018956444329685636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3072,51200,0.06671466430028279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,8192,0.033188445700539485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,1536,5120,0.009920888476901585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,128,7168,0.011592888997660743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,8192,6144,0.02239111065864563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,1536,8192,0.01572355628013611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,2560,0.007179555793603261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,4096,768,0.03746577766206529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,6144,0.007484444313579136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,51200,0.11537244584825303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,128,0.004610666798220741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,128,0.003887999802827835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,128,64,0.007685333490371704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,6144,0.05851288636525472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,128,0.005127111242877113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,8192,0.03443200058407254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,6144,0.05585866504245334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,2560,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,2560,0.04127733243836297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,768,0.009106666677527958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,3584,0.01390044391155243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,65536,2560,0.05058133271005419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,32,0.019778667224778067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,768,128,0.005281777845488654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,3584,0.0069306666652361555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,2048,0.06185422341028849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,10240,0.3518773449791803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,2560,0.07887022362815009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,3584,0.008160889148712158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,4096,0.00629333323902554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,2048,0.36114221149020725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,5120,16384,0.051017777787314526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,2048,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,65536,0.09209244118796454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,3584,0.011696000066068439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,3072,0.057935112052493624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,1536,32,0.002551111082235972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,6144,0.09031111001968384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,12288,7168,5.038503858778212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,12288,8192,0.37705243958367246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,3584,0.022667555345429316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,16384,2560,0.15035643842485216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,6144,0.006505777852402792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,768,0.03409689002566867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,1536,0.06069778071509468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,7168,0.0314648879898919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,5120,5120,0.024452444579866197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,64,0.006573333508438534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,512,0.004898666507667965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,10240,0.07751022444831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,4096,4096,0.12648977173699275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2560,3072,0.016232000456915963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,1024,0.004268444246715969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,12288,128,0.0047271110945277745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,512,0.27849777539571124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,4096,0.008336000144481659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,3072,0.007634667058785756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,512,0.00941955546538035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,6144,0.03816711240344577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,5120,2048,0.00942844483587477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,10240,0.03338311115900675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,3072,0.0092613332801395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,2560,0.0659360024664137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,3584,0.010973333484596677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,3072,0.08584622542063396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,16384,3072,0.2183600001864963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,256,0.020106666617923312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,768,0.041554666227764554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1024,12288,0.2278275489807129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,65536,3.8148824903700085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,32,0.008961777720186446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,4096,0.00927022182279163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,256,0.0033662221911880705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,128,0.00599911105301645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,65536,1.9696701897515192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,3584,0.4707084231906467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,768,0.0046737777690092725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,8192,0.21946310997009277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,51200,0.7071928977966309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,7168,12288,0.6466524336073134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,2048,0.0338986681567298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,51200,2.1984880235460067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,768,65536,0.35920267634921604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,128,0.01124888906876246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3072,3072,0.012262221839692859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,512,0.006235555642180973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,64,0.0057795556883017225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,51200,0.23733954959445527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,64,0.002969777832428614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,1536,0.029658665259679157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,1024,0.03321155574586656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,6144,0.011983999775515662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1024,2048,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2048,6144,0.10057422187593247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,10240,0.03601600064171685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,3584,0.45316802130805117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,128,4096,0.035888890425364174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,3072,0.008079999850855933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,32,0.002784888984428512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,10240,0.02889066603448656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,2048,0.01938577824168735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,16384,0.07846577962239583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,8192,0.023054222265879314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,5120,12288,0.03830844495031569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,512,0.16555733150906032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,256,0.004449777719047335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,768,768,0.003911111089918348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,128,0.010481778118345471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,51200,0.2678719891442193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,7168,3072,0.11004533370335896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,1536,0.006254222244024277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,8192,0.03013333347108629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,256,64,0.0026346666531430352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,8192,0.18444177839491102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,10240,0.44208889537387425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,6144,3072,0.13223111629486084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,12288,6144,0.029996444781621296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,1024,0.6104764408535427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,10240,0.01888177792231242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,2560,0.058865778976016574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,4096,0.031046221653620403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,10240,0.040030221144358315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,512,0.0041333333485656315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,65536,0.15574310885535345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,768,0.05388799972004361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,6144,7168,2.7662150065104165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,128,0.006782222125265334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,512,0.026876444617907207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2560,12288,0.24908089637756348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,4096,0.014765333798196582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,51200,0.10770666599273682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,768,0.04027555717362298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,1536,0.00609777785009808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,1024,0.006969778074158563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,51200,0.05366933345794678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,768,0.007048000064161088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,2560,0.09884266720877753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,3072,0.006670222100284364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,256,0.005388444496525659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,256,0.004396444393528832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,128,2560,0.006270222365856171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,5120,0.02462844385041131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,1536,0.02127555509408315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,10240,65536,1.566362698872884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,64,0.0031840000301599503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,12288,7168,0.05458933446142408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,512,5120,0.012260444462299347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,256,0.004156444635656145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,16384,0.9738293753729926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,4096,0.03907377852333917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,32,0.010951999988820819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,512,0.005177777674463061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,32,0.007600888609886169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,65536,0.14722577730814615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,16384,2560,0.09442044628991021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,12288,0.0796604421403673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,16384,0.26132888264126247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,16384,0.029166221618652344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,6144,0.14693155553605822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,12288,0.14555822478400335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,4096,10240,0.026649778087933857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,2048,0.005899555567238066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,2048,16384,0.11116000016530354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,512,0.04037244452370538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,6144,0.026800000005298193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,5120,0.3370337751176622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,7168,0.012472000386979846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,128,0.003105777833196852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,32,768,0.012791111237472959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,65536,0.12211910883585612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,65536,0.07046222024493747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,1536,0.0099973330895106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,7168,0.5892702208624946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,3072,0.013182222015327878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,2048,0.12160621749030219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,2048,0.029842668109469946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,51200,0.07856533262464735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,2048,0.05790399842792087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,128,0.007560888926188151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,65536,0.11828711297776963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,5120,0.023815999428431194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,1024,0.004488888714048597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,8192,5120,0.021171554923057556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,65536,0.30011733373006183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,1024,0.004960000101062986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,512,0.01937955617904663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2560,16384,0.037830223639806114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,2560,0.019534221953815885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,7168,4096,0.019489778412712943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,1024,0.020502222908867728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,8192,0.18149956067403158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,7168,6144,0.021320889393488567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,3584,0.0233297778500451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,768,0.009284444153308868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,64,0.030172444052166406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,1536,0.02541422181659275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,7168,0.020113777783181932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,32,0.005698666804366642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,7168,0.05582133266660902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,512,0.003625777860482534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,16384,1.576949331495497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,768,4096,0.03764444589614868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,2048,0.009163555171754625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,256,0.0031902222997612427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,8192,0.017510222064124215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,1536,0.006080889039569431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,512,0.008375999828179678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,2048,0.047241777181625366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,256,0.0028604444944196274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,128,0.012010667059156628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,51200,0.05530844463242424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,2560,0.008220444122950235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,2560,0.876007980770535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,7168,0.020804444948832195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,6144,0.023235556152131822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,2048,0.008940444224410588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,5120,0.010696888797812991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,1024,0.018631999691327412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,8192,0.019747555255889893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,10240,0.032815112007988825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,16384,0.027727999620967444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,512,0.005008888741334279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,1024,0.006252444452709622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,51200,0.11497422059377034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,128,0.0052168890833854675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,32,0.005352889084153705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,10240,8192,0.1809991068310208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,51200,10240,0.3397475613488092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,16384,0.25233067406548393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,4096,0.015991111596425373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,128,0.00423200014564726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3072,65536,0.3296639919281006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,3072,0.1426995595296224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,768,0.019845333364274766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,51200,65536,1.8760738372802734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,64,0.003383111208677292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,10240,0.226053343878852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,12288,0.021382222572962444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,3584,0.01977155605951945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,4096,4096,0.4382915496826172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,64,0.012609778179062737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,7168,0.02645422187116411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,65536,0.519244458940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,2560,0.029130667448043823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,128,0.011824888487656912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,3072,0.03071911136309306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3584,1024,0.005655111124118169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,5120,0.2803911103142632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,4096,0.024029332730505202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,8192,0.08340177933375041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,512,0.0053386667536364655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,65536,0.18688800599839953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,64,0.0027599999060233435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,65536,0.44074222776624894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,7168,0.23972177505493164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,4096,0.06135288874308268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,2048,0.006515555497672822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,6144,0.011707555916574268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,16384,0.0581644442346361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,51200,1024,0.02048711147573259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,3072,0.07120711273617215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,512,0.04558488726615906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,16384,0.31255555152893066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,51200,0.13999111122555202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,7168,0.012653333445390066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,1536,0.010889777706729041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,12288,1024,0.048488001028696694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,16384,0.021128889587190416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,2048,0.0075902218619982404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,7168,0.03278755479388767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,65536,10.506601121690538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,5120,0.053427557150522866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,768,6144,0.007928888830873702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,8192,10240,0.03360711203681098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,2048,0.02254044512907664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2048,3584,0.05553333295716179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,16384,8.162674797905817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,4096,0.25319110022650826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,6144,0.17722488774193657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,1536,0.03745422098371718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,256,10240,0.029236445824305218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,7168,0.0347653329372406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,7168,8192,0.23877512084113228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,3072,0.011365333365069495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,768,0.03667555583847894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3584,3584,0.011884444289737277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,512,0.011132444772455426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,512,512,0.0033608888172441055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,32,12288,0.079821334944831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3584,2048,0.019116444720162284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,2560,0.038599110311932035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,6144,128,0.02345422241422865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,12288,512,0.006728888799746831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,51200,0.06221510966618856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,64,0.01774577796459198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,4096,0.04117155406210158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,512,0.043953776359558105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,10240,0.011687111523416309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,256,0.008748444418112436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,16384,0.049250665638181895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,3584,0.00792622235086229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,5120,0.021552888883484736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,12288,0.017464000317785475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,65536,3.1479600270589194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,64,51200,0.3126240041520861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,16384,3584,0.023801777097913954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,6144,0.10965244637595283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,6144,0.03812355465359158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,2048,0.006673777682913675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,64,0.002598222138153182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,2048,0.14376800590091282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,64,0.005150222116046482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,8192,0.019920888874265883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,2560,0.15106844902038574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,4096,12288,0.28472089767456055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,16384,512,0.013776000175211163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,8192,0.03980177640914917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,51200,3584,0.05425599879688687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,128,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,3584,0.006851555572615729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,7168,0.12013778421613906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,2048,0.006984000404675801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,51200,768,0.13905333148108587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,3584,0.008562666674455007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,51200,2048,0.05750311083263821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,7168,0.12702755133310953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,128,0.008074666890833113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,1024,0.0047608887155850725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,7168,0.00776977754301495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2048,32,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,8192,0.06384799877802531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,4096,0.06390666961669922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,128,0.03203999996185303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,512,0.007865777446164025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,768,0.023357333408461675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,4096,0.010427555276287926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,10240,0.2768799993726942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,4096,0.011340444286664328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,6144,0.011724444727102915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,12288,0.029863112502627905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,12288,0.2341217729780409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,1024,0.019725332657496136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,512,0.01202222208182017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,32,0.004311111238267687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,2560,0.022434666752815247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,3584,0.0389048887623681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,256,0.005664000080691443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,2560,0.03107200066248576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1536,32,0.005677333308590784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,1536,0.212682671017117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,65536,512,0.018240888913472492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,128,0.006425777657164468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,1024,0.02183466653029124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,10240,0.020762667059898376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,6144,0.20615733994377983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,51200,1.7846942477756078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,51200,1536,0.36981243557400173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,768,2560,0.008647999829716152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,1536,0.12425600157843696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,65536,1.4983867009480794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,5120,0.01312888910373052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,8192,3584,0.01868977811601427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,65536,0.9624462127685547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,5120,6.457562764485677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,256,0.004034666551484002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,64,0.0034053333931499054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,7168,0.015269334117571512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,2048,0.06493600209554036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,16384,1.0412666532728407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,8192,0.02199911077817281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,1536,0.05624710851245456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,64,0.008221333225568136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,10240,0.036216888162824847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3072,3072,0.011244444383515252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,1536,0.013416888813177744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,2048,0.007094221810499827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,51200,0.994834687974718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,8192,0.014669333895047506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,768,0.007117333511511485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,128,0.0038604442444112566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,8192,0.4952560000949436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,16384,3584,0.05312622255749173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,1536,0.009937778115272522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,51200,256,0.08437155352698432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,768,0.037441776858435735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,1024,0.07695111301210192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,2560,0.015365334020720588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,8192,0.02632444434695774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1536,12288,0.015619556109110514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,512,0.0035786665976047516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,10240,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,3584,0.14660444524553087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,16384,0.7568124665154351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,2048,0.008734222087595198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,256,0.005120000077618493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,128,0.006372444331645966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,3072,0.015976889265908133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,51200,12288,2.9640106625027127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,8192,0.022081777453422546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,256,0.006743999818960826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,256,0.016876444220542908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,3584,0.04537511203024122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,6144,2048,0.08034311400519477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3072,6144,0.10257866647508408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,768,0.0848239991399977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,3072,0.010453333457310995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,8192,0.014136888914638095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,4096,0.008447999755541483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,3072,0.022481777601771887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,512,0.030294222964180842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,2048,0.013553777502642738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,5120,0.30750489234924316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,65536,26.662884182400173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,16384,16384,0.13778221607208252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,128,0.004064889003833135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,16384,2.1004212697347007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,16384,16384,3.2116292317708335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,1536,0.0047733332547876565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,10240,0.025796444879637823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,4096,6144,0.020939555433061387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,10240,0.03303111261791653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2560,8192,0.18343822161356607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1536,12288,0.025455999705526564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,51200,0.0680640008714464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,5120,0.19338577323489717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,10240,0.19665688938564727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,3072,0.01607022186120351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,32,0.06500266657935248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,8192,0.028897778855429754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,1536,0.01424888935354021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,1536,0.037976887491014265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,5120,0.10255555311838786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,7168,256,0.09731200006273057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,6144,0.05538400014241537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,128,0.030477331744299993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,3584,0.009240888886981541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,7168,0.02404533326625824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,8192,0.2421600023905436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,7168,0.01834222177664439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,256,0.0034053333931499054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,5120,512,0.00592533333433999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,8192,16384,0.9886444939507378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,6144,0.011888888974984487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,51200,3072,1.0665217505560982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,768,0.08135644594828288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,7168,2048,0.10832622316148545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,12288,1.108242670694987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,6144,0.029599110285441082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,64,0.005550222264395819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,32,0.0027146666414207886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,7168,4096,0.04970044559902615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,1024,0.009506666825877296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,4096,0.018040888839297824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,512,0.01737422247727712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,64,0.056848002804650195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,4096,0.17344889375898573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,2560,0.05048177639643351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,512,0.04573511083920797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,51200,0.07965066697862414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,65536,1.0748604668511283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,3072,0.02760266595416599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,2048,0.09382488992479111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,8192,0.4102933406829834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,6144,0.017424000634087455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,8192,0.01793333391348521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,768,0.006653333289755716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,128,0.011003555523024665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,10240,0.08321777979532878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,51200,0.06801244285371569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,3584,0.061250666777292885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,6144,0.034635557068718806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,1536,0.0092284447617001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,16384,0.17494222852918836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,10240,0.8280186653137207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,768,0.030410667260487873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,7168,0.025016888976097107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,512,0.013809777796268463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,8192,0.03973599937227037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,128,0.005462222215202119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,128,32,0.0024844445288181305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,5120,0.009425777528021071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,3584,0.009309333231714036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,768,0.0058613332609335584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,3072,0.010584000084135266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,3072,1.3802311155531142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1024,16384,0.022542221678627863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,3584,0.009587555295891231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,16384,0.4443519910176595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,65536,7168,0.23086489571465385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,4096,0.01631644368171692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,128,0.014447111222479077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,768,0.009791110952695211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,12288,256,0.009751111268997192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,4096,0.02696977721320258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,1536,0.12716089354621038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,10240,3072,0.09671911266114976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,6144,16384,0.09204444620344375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,1024,0.016473778420024447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,512,0.005917333480384614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,2048,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,65536,1.0206142001681857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2048,2560,0.0075048887067370946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,2560,0.22363734245300293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,256,3584,0.009556444154845344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,512,0.018026666508780587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,10240,0.030564443932639226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,32,0.02111111084620158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,256,0.006450666735569636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,32,0.003116444374124209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,16384,0.06965688864390056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,12288,0.04745955599678887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,256,0.005631111148330901
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,3072,0.009259555902745988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,2048,0.00674311113026407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,12288,0.024243555135197107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,10240,0.13191200627221003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,10240,6144,0.3408462206522624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,128,0.008164444731341468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,8192,10240,0.05008088880115085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,5120,10240,0.11032178004582723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,6144,768,0.034625778595606484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,8192,8192,0.029057777590221826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,256,51200,0.054584887292650014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,2048,0.013063111239009433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,2560,0.02404622236887614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,2560,0.012318222059143914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,1024,0.02721600068940057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,8192,0.9368551042344836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,51200,17.954549153645832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,1024,2560,0.010308444499969482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,3072,0.009541333549552495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,51200,0.05526489019393921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,128,8192,0.014888000157144336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1536,2560,0.007615110940403408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,51200,0.9629315270317925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,7168,0.16527733537885878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,128,0.0026853332916895547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,256,0.0041831110914548235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,512,0.011890666352377998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,51200,0.1672346724404229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,7168,256,0.0032551110618644287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,128,0.02937244375546773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,2048,0.007542221910423703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,3584,0.1008595559332106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,8192,0.025021332833502028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,2560,0.05946755409240723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,51200,0.056187556849585645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1536,6144,0.0122097780307134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,3072,0.009230222139093611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,5120,0.02550044490231408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,16384,0.053912887970606484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,10240,1536,0.010632889138327705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,64,0.0025946667624844448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,128,0.012228444218635559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,12288,0.03463733196258545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,256,0.014644443988800049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,256,0.022063111265500385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,768,0.010110222631030613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2048,768,0.003936888857020272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,5120,6144,0.02944799926545885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,1536,0.004743111216359668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,2048,0.01956533392270406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,512,256,0.007885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,16384,5120,0.10363022486368816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,16384,0.006041777630647023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,768,0.009522666533788046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,8192,0.006608888920810487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,6144,0.04043822156058417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,1536,0.005528000079923206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,3072,0.05390666590796577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,32,1024,0.004504000147183736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,12288,8192,0.07708799839019775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,4096,0.025949334104855854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,7168,8192,0.339105765024821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2048,6144,0.14264267020755345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,7168,0.023265777362717524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,5120,3072,0.010240889257854885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,32,0.005135111096832487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,5120,512,0.003641777982314428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,128,512,0.01311644415060679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,65536,0.1055235531595018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,3072,0.016306666864289176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,4096,0.013751111096805997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,128,0.005909333212508096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,512,0.4791955418056912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,16384,51200,0.25594133800930446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,2048,4096,0.051810665263069995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,8192,16384,0.20206577248043486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,64,0.003055111194650332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,2560,0.10333333412806193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,3584,0.019095111224386428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,16384,0.03371466530693902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,65536,0.0774844421280755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,10240,4096,0.01846577723821004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,3584,0.6454959975348579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,7168,512,0.021015110943052504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,51200,10240,3.2090835571289062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,512,0.0118222220076455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,32,51200,0.07392888598971896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,4096,0.022192888789706763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,16384,1.6216977437337239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,5120,3072,0.08519022332297431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,256,0.0032115555885765287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,65536,6144,0.2518400086296929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,3072,2560,0.0402728882100847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,3072,0.010114666488435533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,2048,0.03370044297642178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,3584,0.2772408856285943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,65536,0.08945866425832112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,7168,0.7916186650594076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,4096,0.021891555852360193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,3072,0.05039733317163256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,2560,0.01860000027550591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,1536,0.01332800007528729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,768,0.025236444340811834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,768,0.012979555461141797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,8192,2048,0.04836888776885139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,32,16384,0.01847111185391744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3072,3072,0.013552889227867126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2560,32,0.002785777880085839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,768,256,0.004063999901215236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,3072,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,65536,0.2589884334140354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,5120,768,0.012663111090660095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,4096,128,0.003144888828198115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3072,64,0.006385777973466449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,7168,0.011092444260915121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,128,0.01812977757718828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,10240,0.14628533522288004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,128,0.0028533333291610083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,4096,1.200850698682997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,768,0.006958222223652734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,65536,8192,0.29101954566107857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,4096,256,0.006526222245560751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,5120,0.011902222202883827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,32,0.010300444232092964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,512,0.006311999840868845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,128,0.0064560001095136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,16384,256,0.0522382226255205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,10240,51200,4.530968136257596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,256,3072,0.00868266655339135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,7168,0.013824000126785703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,256,0.028499556912316218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,10240,0.038375112745496966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,1024,0.006120000034570694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,64,0.007947555846638149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,51200,1536,0.026916444301605225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,7168,0.018924444913864136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,2048,0.03468888998031616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,1024,0.005535111245181825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,51200,1.4018835491604273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,12288,0.5888497564527724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,12288,64,0.021653332644038733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,5120,0.053094221485985645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,32,0.033412443266974554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,256,0.003206222214632564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,2048,0.023599111371570166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,12288,0.20744266774919298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,128,0.02000266644689772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,768,10240,0.11303733454810248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,51200,0.22505601247151694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,128,7168,0.06570755773120456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,51200,0.02572000026702881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3072,10240,0.12551821602715388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,2560,512,0.026318222284317017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,65536,2560,0.05001599921120537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,10240,0.03691822290420532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,12288,0.018431999617152743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,3584,0.009906666974226633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,51200,0.12017156018151177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,1024,0.009364444348547194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,3072,0.012661333713266583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,64,7168,0.006531555619504716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,1536,3584,0.052240888277689614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,12288,0.057518223921457924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,32,0.005072888814740711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,65536,51200,2.5846676296657987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,65536,0.367595566643609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,65536,32,0.12477866808573405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,10240,0.10345600048700969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,256,0.007227555745177799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,8192,3072,0.7248684565226237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,10240,0.42584888140360516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,65536,0.2050328916973538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,8192,0.10897333092159694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,256,0.004816888935036129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,2560,0.007270221908887227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,6144,0.02163200080394745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,2560,0.005681777579916849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3584,3072,0.010973333484596677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,64,0.020464888877338834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,512,0.006265777680608962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,768,0.009369778136412302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,6144,0.0158897770775689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,768,0.004933333231343163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3584,6144,0.19961867067548963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,4096,27.0464104546441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,2560,0.08454400300979614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,32,0.0075288886825243635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,51200,10240,3.9669715033637156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,65536,3.071411556667752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,7168,0.4005120065477159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,6144,0.02684088879161411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,10240,5120,0.3869484530554877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,5120,2048,0.12335556083255345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,768,0.008225777910815345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,8192,12288,0.038508445024490356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,10240,0.5305253134833442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,32,0.017080000705189176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,51200,2.7149136861165366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,2560,0.010600888894663917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,3072,0.01868977811601427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,7168,0.010192888478438059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,768,3584,0.03593511051601834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,3072,0.007086222370465596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,16384,0.07885333564546373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,64,2560,0.006027555714050929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,2560,0.007740444607204861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,2560,64,0.005094222310516569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,2560,12288,0.01830400029818217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,10240,32,0.030108445220523413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,10240,3584,0.02984444300333659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,16384,0.030999110804663763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,3072,0.015501333607567681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,51200,512,0.20445688565572104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,32,0.0069653333889113525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,128,256,0.003314666656984223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3072,256,0.004457777573002709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,768,0.007846222155623967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,128,0.004293333325121137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,65536,2.0835653940836587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,4096,0.01979377700222863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,3584,0.017954667409261067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,3584,0.010409778190983666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,1024,0.008630221916569604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,51200,2.7141840193006725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,128,0.00453244439429707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,5120,0.046642667717403836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,65536,10240,0.18363377783033583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,7168,0.021532444490326777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,5120,0.09773600101470947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,16384,14.3753170437283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,8192,0.08950310945510864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,64,0.003999999827808804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2048,16384,0.2657119962904188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,256,0.010335110955768162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,3584,0.11448800563812256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,4096,0.012690666649076672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,7168,0.02602400051222907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,7168,0.006380444599522485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,5120,0.0700026684337192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,768,0.09932621982362534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,7168,0.027218666341569688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,5120,8192,0.07296266820695665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,1536,0.0063386667105886675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,128,0.005421333428886201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,10240,512,0.04155733187993368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,128,0.004884444591071871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,16384,3584,0.022942221826977197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,51200,6144,0.09381244579950969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,3072,0.008652444514963362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,8192,0.3640444543626573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,256,10240,0.01348533315791024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2560,256,0.004693333473470476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,128,0.008979555633332994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,2560,0.32687200440300834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,51200,0.09043200148476495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,5120,5120,0.042640888028674655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,8192,128,0.006237333433495627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,32,0.003128000017669466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,65536,51200,1.7454355027940538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,6144,0.118724438879225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,768,0.01331822243001726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,3072,8.574698554144964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,6144,0.08282133605745104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,12288,0.0184542222155465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,8192,0.015995555453830294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,256,0.01074044406414032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,6144,0.20713067054748535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,51200,0.07256444295247395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,512,0.16949600643581816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,1024,0.056610663731892906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,768,0.0057760001056724125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,32,6144,0.00981155534585317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,768,1536,0.012812444733248817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,16384,0.00639111093348927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,32,0.012714666624863943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,512,0.010101333260536194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,5120,0.0465866658422682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,2048,0.01830400029818217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,2048,0.06887822018729316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,1536,0.007122666471534305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,3072,0.006455111006895701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,1024,0.008184000021881526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,16384,0.5258311165703667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,256,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,1536,0.045048889186647206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,5120,0.03482222225930955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,256,256,0.00738844441043006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,5120,0.017959111266665988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,8192,0.03937955697377523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2048,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,6144,51200,0.1457537810007731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,512,65536,0.06622311141755846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,1024,0.4856239954630534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,6144,0.6007564332750108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,6144,0.1818915605545044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,4096,0.0058444444504049085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,3584,0.09850311279296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,16384,0.062367108133104115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,64,0.0111191107167138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,10240,0.044822222656673856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,12288,51200,3.7558826870388455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,3584,0.07618400123384264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,3072,0.616354677412245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,2560,0.3493368890550401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,3072,0.009904888768990835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,768,0.005672889037264719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,3584,0.024567110670937434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,65536,4096,0.1338097784254286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,51200,0.3891022205352783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,65536,0.4793039957682292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,51200,512,0.12774044937557644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,512,0.009129777550697327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,1024,0.006762666834725275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,10240,3584,0.019680889116393197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,512,0.003826666623353958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,2560,0.025900445050663416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,3584,0.1303902202182346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,768,3072,0.0073902226156658595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,65536,0.20583111710018584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,256,0.00555377784702513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,64,0.032609777318106756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,65536,3584,0.3785155614217122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,51200,0.03573066658443875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,4096,7168,0.015419556034935845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,512,0.008281778130266402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,2048,0.008912888665994009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,65536,5120,1.1975573433770073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,3072,0.0074471111098925276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,128,0.024352888266245525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,65536,3072,1.1996355056762695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,5120,0.02827199962404039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,1024,0.13319378428988987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,10240,1536,0.12130488289727105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,1536,0.01644711030854119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3072,4096,0.016319110989570618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,4096,0.010455999937322406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,6144,0.03306844499376085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,5120,0.010803555448849997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,10240,0.09541510873370701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,3072,0.8228355513678657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,7168,0.014689778288205465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,16384,6.172194586859809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,768,0.039868444204330444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,5120,0.15344266096750894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,12288,0.05201155609554715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,65536,6144,1.1403510835435655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,16384,21.090042114257812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1536,2560,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,2048,0.031962666246626116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,5120,12288,0.4863226678636339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,8192,0.07794488800896539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,2048,0.6143724653455946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,10240,0.05051999953058031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,1536,0.008101333346631791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,1024,0.006320000108745363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,3072,8192,0.02013777858681149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,1536,12288,0.04423911041683621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,256,6144,0.012720000412729053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,128,2048,0.00828888929552502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,51200,0.7826142311096191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,2560,0.013376889129479727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,65536,51200,0.986413319905599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,5120,3584,0.012869333227475485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,5120,256,0.08440622356202866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,3072,0.03331289026472304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,768,0.008246222303973304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,1536,0.04467644294102987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,2048,0.053463111321131386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,51200,1.2940560446845162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,16384,12288,0.11430488692389594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,4096,0.009559111462699043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,128,0.006486222147941589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,64,0.07882489098442926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,3072,0.027504889501465693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,8192,0.05004355642530653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,1024,0.24948443306816948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,128,0.006196444647179709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,1536,0.004829333474238713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2048,51200,0.8299902280171713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,4096,0.05150666501786974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,256,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,1024,0.4765644603305393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,8192,10240,0.04128799835840861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,51200,0.06053688791063097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,2048,0.011284444067213269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,256,0.024990222520298425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3584,10240,0.028811554114023846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,512,0.005562666803598404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,512,0.01627200014061398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,12288,5120,0.025921778546439275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2560,768,0.027540443672074214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,512,8192,0.03791555431154039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,2560,0.3477235635121663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,65536,1.088921758863661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,7168,0.2163297732671102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,12288,4.852531433105469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,256,0.007260444263617198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,5120,0.01847555571132236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,51200,1024,0.038859556118647255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,1024,0.011367999845080905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,10240,0.2127164469824897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,7168,0.2802693314022488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,256,0.0032133333798911837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,3072,0.008608889248636033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,256,256,0.005618666609128316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2560,32,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,65536,16384,6.019647810194228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,51200,30.722646077473957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3072,1536,0.010101333260536194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,4096,0.07086755832036336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,32,0.011046222514576383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,12288,128,0.03493155704604255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,8192,4096,0.11666311158074273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,16384,0.0888951089647081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,12288,0.08951289123959011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2048,128,0.005007110950019625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,3584,0.04344800114631653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,128,0.002854222224818336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,6144,0.006592888798978593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,12288,0.007167999943097432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,512,1536,0.005721777677536011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,65536,1.7589662339952257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,3584,0.011428444749779172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,768,0.007777777810891469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,1024,0.0041662222809261745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,768,256,0.042862223254309766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,512,0.07550044192208184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,6144,0.397140449947781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,6144,0.338831106821696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,16384,768,0.03677155574162801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,4096,51200,0.13767377535502115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,16384,0.08221866687138875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,4096,0.021007110675175984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,256,768,0.08431022034751044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,64,0.005883555445406172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,16384,0.04826400015089247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,5120,10240,0.032837334606382586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,10240,0.15222133530510798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,1536,0.035828444692823626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,7168,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,51200,128,0.13980178038279215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,5120,0.02424533334043291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,3584,0.3495937718285455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,65536,0.15926044517093235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,12288,8192,0.20666755570305717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,3072,0.022664000590642292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,5120,0.01038755559258991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,4096,0.013530666629473368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,1536,0.014754666222466363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,512,51200,3.241369883219401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,10240,0.03515999846988254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2560,65536,0.5131128629048666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,512,0.02086755633354187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,512,7168,0.03832622369130453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,3072,0.00884622252649731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,5120,0.018544000056054857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,10240,0.08671377764807807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,3584,0.022522666388087805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,5120,768,0.006250666661394968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,2048,0.03538844320509169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,8192,1536,0.009479111267460717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,768,4096,0.015021334091822306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,6144,0.02628355556064182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,8192,0.21368088987138537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,4096,7168,0.01628444426589542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,1024,0.005943999936183293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,64,0.0026604444202449587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,10240,0.026759111218982275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1024,1536,0.0063768890168931745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,32,0.006568000134494569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,1024,0.006029333505365584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,2048,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,32,10240,0.023401778605249193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,1024,7168,0.012947555217478009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,5120,0.0103502223889033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,5120,0.1397591167026096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,5120,0.008987555901209513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,256,5120,0.009963555468453301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,4096,0.009289777941173976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,256,0.014744000302420722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,8192,1536,0.015757333901193406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,7168,0.012483555409643384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,128,0.013030221892727746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,256,7168,0.061985777484046094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2560,5120,0.01406222250726488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,2560,64,0.012164444559150271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,10240,0.3127635584937202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,5120,0.034895999564064875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,5120,51200,2.3795973459879556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,256,0.0052897776994440295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,12288,0.0520684454176161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,12288,0.2775599956512451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1024,16384,0.058014220661587186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,4096,7168,0.031269331773122154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,512,0.004864000197913912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,65536,3072,0.564008871714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,4096,0.029083556599087183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,3072,0.012497777740160624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,6144,0.2690746784210205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,5120,0.01697688963678148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,32,512,0.010259555445777046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,768,0.008433777425024245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,1536,0.019161777363883126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,5120,0.29626133706834584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,2048,0.17990310986836752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,12288,0.019851555426915485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,1536,0.006451555424266391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,1024,0.20338755183749727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,768,0.07638933261235555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,65536,6144,10.66483900282118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,1024,0.018452444010310702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,2048,1.3035288916693795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,1024,0.06070578098297119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,8192,0.029185778564876978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,65536,4.745079888237847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,7168,51200,0.11106755336125691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,3072,0.2090391053093804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,3584,0.009256000320116678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,6144,0.11062133312225342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,10240,1536,0.20401955975426567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,51200,0.5078978008694118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,51200,5.274368709988065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,768,0.006994666324721442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,16384,0.11034044292238022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,3072,0.0629404452111986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,4096,0.8384942478603787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,1536,0.008325333396593729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,5120,0.012668444050682915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,2048,0.014480888843536377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,12288,0.023423999547958374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,8192,0.00777333312564426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,256,0.006298666612969504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,12288,4096,0.029121776421864826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,10240,1536,0.011759111450778114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,7168,0.011024889018800525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,51200,0.03251466817326016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,4096,12288,0.1332168844011095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,10240,0.4500480228000217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,128,3072,0.01094133324093289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,128,0.003370666669474708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,128,0.013088888592190213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,768,0.027569777435726587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,65536,10240,3.6115743849012585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,256,3072,0.010816888676749336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,65536,0.08281866709391277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,65536,0.023224888576401606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2048,10240,0.016563556260532804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,512,0.0057839999596277876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,5120,3072,0.031368888086742826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,64,0.0026791110220882627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,51200,32,0.1495751142501831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,3584,0.041399998797310725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,16384,0.03444088829888238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,6144,0.01406222250726488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,32,0.010984888507260216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,512,0.004295111116435793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,1536,0.008153777983453538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,12288,0.22374844551086426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,2560,0.006017777654859755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,6144,0.16862932840983072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,6144,0.011623111036088733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,1024,0.005790222022268508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,6144,0.20042310820685494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,16384,12288,0.16388711664411756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,65536,8.288675944010416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,8192,0.06898844242095947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,65536,0.10964533355500962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,16384,0.06393066379759046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,2048,0.009725333087974125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,2048,0.01180355581972334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,5120,5120,0.016931555337376065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,8192,0.5718097686767578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,8192,0.03818488783306546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,3584,0.017453332742055256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,8192,0.02086399992307027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,1024,0.02456533412138621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,65536,8192,29.120768229166668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,3072,0.010764444039927589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,51200,1.084935082329644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1024,32,0.0038195554580953387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,5120,0.01420088940196567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,8192,0.008557333714432186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,128,51200,0.6198906898498535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,65536,5.534138573540582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,64,8192,0.052095999320348106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,16384,128,0.018561777141359117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,8192,0.05261066555976868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,64,0.005881777654091517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,7168,0.016526222229003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,10240,0.017468444175190397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,1024,0.02360444433159298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,768,0.017337777548366122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,768,6144,0.07039378086725871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1536,2048,0.02777066661251916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,2048,0.009900444083743626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,768,0.028953777419196233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,3584,0.028304888142479792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,4096,0.12490133444468181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3072,8192,0.01926666663752662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,6144,0.006242666807439592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,12288,0.9815102683173286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,4096,0.04868533213933309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,3072,0.008105778031879002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,5120,0.012804444465372296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,16384,0.08953333563274807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,128,0.003456888927353753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,512,0.005147555636035072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,128,0.005145777844720417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,3584,0.02713955607679155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,12288,0.015578667322794596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,10240,0.031197332673602637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,64,1024,0.011715555356608497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,32,0.0188711103465822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,64,0.006340444501903322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,1024,0.05920000208748711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,256,10240,0.13260800308651394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,1536,0.008681777450773451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,7168,0.021645334031846788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,768,0.004616888860861461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,7168,32,0.005565333283609814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,768,0.015849777393870883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,768,0.005111999809741974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,2048,0.01310755560795466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,2048,0.052240888277689614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,64,5120,0.011847111086050669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,1536,0.005392000079154968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,4096,0.014541332920392355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,1536,0.03675555520587497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,256,2560,0.006646222124497096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,256,0.029096000724368628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,65536,0.27145422829522026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,65536,12288,3.9665421379937063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,32,0.008534222013420528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,6144,0.2927244504292806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1024,1024,0.010608000059922537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,16384,0.06041777796215481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,10240,0.04011822077963088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,10240,2048,0.07631999916500516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,12288,0.016396444704797532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,16384,0.02349866595533159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,10240,0.059169775909847684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,16384,0.16546577877468532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,32,0.008184889124499427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,7168,0.02812622321976556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,256,2048,0.005894222193294101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,2560,0.012774222426944308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,3584,0.017647110753589206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,12288,0.41820621490478516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,4096,0.00923644420173433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,12288,0.041386667225095965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,64,51200,0.05385955505900913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,512,32,0.005503111001518037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,4096,0.011178666518794166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,8192,6144,0.09069689114888509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,512,0.007215110792054071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,256,2560,0.008848889006508721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,128,0.003999111139112049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,7168,0.0277146663930681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,2048,0.006446222050322427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,128,0.2278204494052463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,12288,0.04649244414435493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,12288,128,0.0046737777690092725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,16384,0.5486613379584419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,1536,0.030283557044135198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,512,0.008810666700204214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,8192,0.009182222187519073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,7168,0.7650986777411567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,64,0.008124444219801161
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,128,8192,0.012561777399645912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,768,768,0.012372444073359171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,6144,0.009773333039548662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,8192,0.017487999465730455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,4096,0.03775644302368164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,6144,0.02699111070897844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,768,0.10507822036743164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,512,0.004481777962711123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,256,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,2560,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,6144,0.01806311143769158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,51200,1.3316284815470378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,7168,768,0.04363377888997396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,7168,0.7743520206875272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,1536,0.03269155489073859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,5120,65536,0.1308008829752604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,4096,0.02384088933467865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,51200,768,0.026390223039521113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,768,0.012435555458068848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,10240,7168,0.02881866693496704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,12288,2048,0.014087999860445658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,64,0.003313777761326896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,12288,0.2564844555324978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2560,512,0.003448888866437806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,768,0.00452622233165635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,1536,0.00885511106914944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,128,3584,0.02008888953261905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,256,0.0028497777465316984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,65536,1.900485356648763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,7168,12288,0.1458266708585951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,10240,0.007263999846246507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,7168,2560,0.014359111587206522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,16384,0.006350222147173352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,2048,0.37107377582126194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,51200,0.032305777072906494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,64,0.012441777520709567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,3584,0.11671022574106853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,32,6144,0.0402817792362637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,2560,0.010978666444619497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,7168,0.012409778104888068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,8192,0.6220871077643501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,10240,65536,0.5619040065341526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,51200,0.03493955400254991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,7168,2.5879323747422958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,5120,0.07858755853441027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,51200,1.018705791897244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,512,0.014241778188281588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,2560,0.007157333195209503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,1024,0.10571822192933823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,2560,0.0861733357111613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,10240,0.28041421042548287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,3072,512,0.017942221628295053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,64,0.010059555371602377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,128,0.0038542221817705366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,2048,0.00675999994079272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,5120,0.008634666601816813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,8192,0.016626665989557903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,8192,1.766550170050727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,1024,0.004990222139490975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,51200,0.10763821999231975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,6144,0.037462221251593694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,512,64,0.0029519999192820657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,768,0.0036755556033717263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,7168,2560,0.015579554769727917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,3072,0.033039109574423894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,7168,4096,0.2021155489815606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,128,0.007673777639865875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,6144,0.020415110720528495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,64,0.019205333458052743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,7168,0.022641777992248535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,768,0.004127999974621667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,7168,12288,0.26289600796169704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,128,6144,0.023860444625218708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,512,0.015936889582210116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,128,0.004998222407367495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,65536,1.2861999935574002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,32,0.006067555397748947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,1024,0.3068399959140354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,51200,0.16516533162858751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,1024,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,64,0.002875555513633622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,3584,0.24902576870388457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,12288,0.055890666113959424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,5120,0.01945333348380195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,4096,0.14600711398654515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,3584,0.014856888188256158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,12288,0.053470220830705434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,5120,0.22283466657002768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,4096,2048,0.012326222327020435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,5120,0.008992888861232335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,65536,0.26411644617716473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,12288,0.04604533314704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,4096,0.012502222425407834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,51200,1536,0.027183110515276592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,3584,0.03077600068516201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,16384,0.03547999925083584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,51200,0.19016622172461617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,12288,3072,0.21588444709777832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,5120,0.020170667105250888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,64,0.0027173333283927706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,1024,0.03713333275583055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,4096,0.6750062306722006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,12288,65536,3.8708398607042103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,51200,8192,0.6056915389166938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,128,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,768,0.008004444340864817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,64,3584,0.0107022225856781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,3584,0.1030835575527615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2048,1024,0.004173333446184794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,12288,0.05531644490030077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,10240,512,0.04495111107826233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2048,51200,0.0603297750155131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,2560,0.5699671109517416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,5120,0.02160800000031789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,4096,2560,0.7129706806606717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,256,0.00554666668176651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,7168,0.01827911039193471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,3072,0.04699110984802246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,256,0.005324444423119227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,512,0.015511999527613321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,2048,0.026296888788541157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,16384,51200,0.38979821734958225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,8192,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,16384,0.10086844364802043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,1024,0.004230222354332606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,512,0.02239022155602773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,3072,0.007716444631417592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,512,8192,0.05381688806745741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,256,0.004679999831649992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,51200,0.05734311209784614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,12288,0.031126221021016438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,51200,0.666776869032118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,64,0.006096000058783426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,1536,0.005030222237110138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,8192,0.007044444481531779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,3584,0.013620444469981723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,10240,1.271609730190701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,8192,51200,4.188442654079862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,6144,2048,0.014501333236694336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,2048,0.04046666622161865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,8192,0.3030800024668376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,2560,0.22346133655971953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,12288,0.017247110605239868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,10240,0.017145777742067974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,2048,0.01262399968173769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,128,0.019488889310095046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1024,4096,0.008238222036096785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,8192,128,0.020449777444203694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,51200,1.8593777550591364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,2560,0.04145955708291795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,3584,0.09418311383989121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,6144,0.0173688895172543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,7168,0.22098222043779162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,12288,0.02644444505373637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,10240,128,0.036863999234305486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,32,0.005883555445406172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,4096,128,0.0122097780307134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,768,12288,0.021879111727078755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,1536,0.04345600141419304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,6144,128,0.01463822192615933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,2560,0.05279111199908786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,8192,0.07283733288447063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3584,51200,0.09224622117148505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,6144,0.016153777639071148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,2048,3.6813341776529946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,51200,1.350431972079807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,64,0.0063279999627007385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,16384,0.05042044321695963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,256,0.012870222330093384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,65536,1536,0.03626755542225308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,65536,10240,0.19750400384267172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,2048,0.00499199993080563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,16384,0.04859644505712721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,128,6144,0.01388177772363027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,5120,0.1354666683408949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,7168,0.01809866726398468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1536,1536,0.03631555702951219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,65536,0.17480177349514434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,6144,0.17339111698998344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,2048,0.0572320024172465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,12288,7168,0.5355724228752984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,4096,128,0.02622844444380866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,8192,0.019574221637513902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,2048,0.007527111305130853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,7168,8192,0.0336480008231269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,128,128,0.006171555568774541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1024,32,0.005255110975768831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3584,256,0.11521866586473252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,7168,0.18713333871629503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,6144,768,0.008915555146005418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,4096,7168,0.17559733655717638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,32,64,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,256,2560,0.006329777754015393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,1024,0.0058133333093590206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,64,0.005018666800525454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,64,512,0.0031235555393828284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,16384,0.29172444343566895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,2560,0.006659555352396435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,1536,0.01645777788427141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,128,0.006136888845099344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,512,5120,0.010025777750545079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,1536,0.023367111881573994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,4096,0.06598489152060615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,3072,3584,0.0800044470363193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,12288,8192,0.05335022343529595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,128,0.004204444587230682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,2048,14.250544230143229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,5120,1.6184906429714625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,4096,7168,0.18867466184828016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,3072,0.010536889235178629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,64,0.004109333372778363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1024,128,0.003915555361244413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,3072,0.20757244692908394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,4096,3072,0.041174223025639854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,3584,65536,0.12106577555338542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,6144,128,0.011665778027640449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,4096,1536,0.00816177742348777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,7168,7168,0.025840888420740765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,4096,0.2837866677178277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,3584,0.3112240102556017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,51200,0.13576177755991617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3072,3072,0.009974222216341231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,1536,0.00519466648499171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,3072,0.009692444569534725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,32,0.007511110769377814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,2560,0.019474666979577806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,6144,0.11216178205278184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,2560,0.011203555597199334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,256,0.009296000003814697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,256,0.009521777431170145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,768,0.007197333706749811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,5120,0.06700533628463745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,5120,0.013572444518407186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,8192,0.02179466684659322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,51200,32,0.08270755741331312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,6144,2048,0.054195556375715465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,3584,0.017119111286269296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,128,3072,0.008435555630260045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,4096,0.014458666245142618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,65536,0.07409955395592584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,512,0.0037475555307335327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,2560,0.01610488858487871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,12288,0.2106106678644816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,128,0.013614222407341003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,12288,0.0177839994430542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,3584,0.033365332418017916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,128,4096,0.014154665999942355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,10240,0.006620444357395172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,5120,0.008584000170230865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,768,0.04309066798951891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,512,0.008173333273993598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,1024,0.006544000158707301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,16384,0.023189332750108507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,7168,0.19637155532836914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1536,2048,0.32435109880235463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,51200,0.42310134569803876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,768,12288,0.05252266592449612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,8192,12288,0.0683386656973097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,128,4096,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,12288,8192,0.6867173512776693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,7168,0.2395706706576877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3072,8192,1.8320293426513672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,3072,0.019736889335844252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,4096,0.008934222161769867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,2048,0.008731555607583787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,1024,2048,0.007648888561460707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,256,7168,0.012697777814335294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,4096,4096,0.07249066564771864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,128,0.004421333471934001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,8192,0.01161688897344801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2560,65536,0.07291822301016913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,10240,0.4871795442369249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,51200,0.15257333384619817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,6144,0.020523554748959012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,65536,0.14295821719699436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,65536,512,0.22265333599514434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,256,1024,0.005201777650250329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,1024,0.00369777778784434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,12288,1536,0.13021955225202772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,65536,1.525025791592068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,3584,0.01872711049185859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,768,0.014362666341993542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,4096,7168,0.026919111609458923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,6144,0.020544888244734872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,5120,0.024911110599835713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,16384,0.09113066726260716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,768,768,0.003882666842805015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,64,5120,0.006478222293986215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,512,0.008110221889283922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,512,0.005628444254398346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,7168,2560,0.013757333159446716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,5120,0.03935022155443827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,65536,2.022909376356337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,6144,16384,0.06938755512237549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,1536,0.060361776087019176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,3584,3.4048692915174694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,16384,1.1104639901055229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,2560,768,0.00686488880051507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,65536,1536,0.03814311159981622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,256,0.003954666770166821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,1024,0.007391999993059371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,8192,7168,0.3806017769707574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,16384,0.06921155585183038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,768,0.044864889648225575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,1536,0.01240799989965227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,7168,0.009248889154858058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,10240,0.023846222294701472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,12288,0.04561422268549601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,10240,0.2005688879224989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,2048,0.2364284462398953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,256,0.011009777585665384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,16384,0.6863297886318631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,6144,0.16712444358401826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,7168,512,0.009208888643317753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,3584,1.1256115171644423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1024,10240,0.013786666923099093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,51200,1.2196168899536133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,2048,0.013635555903116861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,2560,0.015856888559129503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,512,0.005476444545719359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,3072,0.01125955581665039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,65536,0.15360532866583929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,512,0.0032711111836963226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,8192,2048,0.013883555928866068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,6144,6144,0.020012444920010038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,65536,0.015027556154463025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,7168,0.5299004448784722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,64,0.015314666761292351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,5120,0.0075288886825243635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,256,0.0034231110993358824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,3584,0.020447999238967896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,51200,0.6899289025200738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,5120,0.9600701861911349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,5120,16384,0.6699920230441623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,8192,1024,0.011183111204041375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,12288,0.04385777645640903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,128,0.00832177781396442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,512,51200,0.1819982263776991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,5120,0.02183377742767334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,1536,0.05341688791910807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,2048,0.006904888898134232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,768,0.009276444713274637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,256,0.024625778198242188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,256,0.006263111200597551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,2560,0.023665777511066858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,512,0.0037466668420367767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,64,0.002603555512097147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,65536,0.34809690051608616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,51200,0.10042666726642185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,51200,0.23193955421447754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,2048,0.007256888680987888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,768,1024,0.0058764442801475525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,65536,3.583001030815972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,64,3584,0.011001778145631155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,12288,0.45457156499226886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,16384,10240,0.7561608950297037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,256,0.004360888981156879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,51200,17.327746921115452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,4096,1536,0.008408000071843466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,512,0.005004444470008214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,3584,0.05369244350327385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,64,65536,0.44142844941880965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,16384,0.4687360127766927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,5120,3072,0.09038755628797744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,7168,0.08984711435106064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,768,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,7168,0.0343137780825297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,6144,0.16414666175842285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,2048,0.006180444525347815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,6144,1.0268800523546007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,16384,1.5090612835354273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,64,0.0033093332830402586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,128,0.003552888830502828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,8192,16384,0.0628933310508728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,256,0.008972444468074376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,3584,0.02562933332390255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,12288,0.047741333643595375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,5120,0.012493333054913415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,2560,0.14581778314378527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,7168,0.01511111193233066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,768,0.05207822057935926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,768,0.0036355555057525635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,512,7168,0.011138666835096149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,1024,0.004774222357405556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,768,0.014183999763594734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,7168,0.01401422255569034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,5120,0.013056000073750814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,1536,0.05288533369700114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,128,0.007431999676757389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,256,0.003072000005178981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,1024,0.008185777399275038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,2048,0.10978221893310547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,4096,12288,0.30743198924594456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,5120,0.6345155504014757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,256,0.08003644148508708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,12288,0.06727021932601929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,16384,0.07345333364274767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,5120,0.1867920027838813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,65536,2.368377685546875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,65536,4096,0.1367573340733846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2048,3072,0.3782186773088243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,7168,7168,0.026271111435360376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,65536,0.07301066981421576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,65536,4.76494386461046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,64,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,51200,0.1697617769241333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2560,3584,0.009903999666372934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,16384,64,0.011944000091817645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,2048,0.05225155419773526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2048,16384,0.03755377729733785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,5120,0.16924088531070283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,1024,1024,0.006024000131421619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,8192,0.020244444410006206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,1536,0.0746142201953464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,12288,0.45304086473253036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,4096,0.02181511123975118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,2048,0.021235555410385132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,512,0.010972444381978778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,65536,0.03472000029351976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,512,0.0052133335007561584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,256,0.01704355577627818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2560,65536,0.07308533456590441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,2048,0.07167200247446696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,51200,0.07826933595869276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,128,0.004056888735956616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,7168,0.013543999857372709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,3584,0.010843555960390302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,16384,0.4688408639695909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,16384,2560,0.07039555576112536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,8192,256,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,51200,5120,0.15407200654347739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,51200,0.0992471112145318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,8192,6144,0.024855999483002558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,16384,12288,0.06966488891177706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1536,1024,0.005137777576843898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,2560,0.007698666718271043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2560,65536,0.09404622183905707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,10240,7168,0.046599109967549644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,16384,0.27315910657246906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,12288,0.04495822058783638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,4096,0.0792106654908922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,768,0.015899555550681222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,512,0.03174133433236016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,2048,0.007066667079925537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,16384,5120,1.0791635513305664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,8192,4096,0.1823422246509128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,512,0.43426044782002765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,512,0.0124177775449223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,4096,12288,0.036031110419167414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,256,0.005043555465009478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,256,0.2258080111609565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,16384,0.052780446079042226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,5120,10240,0.02270577847957611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,6144,0.4427351156870524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,16384,0.026919111609458923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,10240,0.5699111090766059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,6144,128,0.0037600000699361167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,768,0.005040888984998067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,8192,0.23964977264404297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,512,0.00600622221827507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,51200,0.26018844710456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,3072,0.009294222626421187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1536,256,0.0058702222175068325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,1024,0.011504000259770287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,7168,2.9000701904296875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,16384,0.05228622092141045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,256,0.011310222248236338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,10240,0.08024799823760986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,16384,2560,0.02418400016095903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,7168,0.01610133383009169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,7168,0.024783111280865137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,10240,0.04632800155215793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,65536,7168,0.1331004434161716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,3584,0.08499822351667617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,1024,0.006593777901596493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,12288,0.019613333874278598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,512,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,128,0.002833777831660377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,12288,0.03722044494416978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,2560,0.007185777856243982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,10240,65536,0.19232000244988334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,2560,0.7091982099745008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,64,0.005163555757866966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,16384,0.428908454047309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,10240,0.2788622114393446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,1536,51200,0.04756177796257866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,128,7168,0.017416889468828838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,2560,0.00921688891119427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,1024,0.0053591111467944244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,256,0.010107555323176915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2048,128,0.005490666462315454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,32,0.005992888990375731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,768,256,0.01903288894229465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,12288,1.1496648788452148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,6144,6144,0.44061867396036786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,1536,0.01387555566098955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,65536,1.0831022262573242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,8192,64,0.01204711116022534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,5120,10.652688768174913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1024,5120,0.016348444753222995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,3584,0.1336559984419081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,64,7168,0.04201333390341865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,2048,0.011523555550310345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,8192,5120,0.01964977714750502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,2048,0.008340444829728868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,768,0.005168888717889786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,12288,0.015389333168665567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,4096,2048,0.06507200002670288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,8192,0.014192000031471252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,2560,0.016951110627916124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,6144,0.010821333361996545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,3072,0.010810666614108615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,51200,128,0.010806221928861407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,16384,0.24913422266642252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,3584,0.017307556337780423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,128,12288,0.03120888935195075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,16384,0.08060889111624824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,12288,0.11963821782006158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,768,0.003765333443880081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3072,4096,0.010658666491508484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,7168,0.01755199995305803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,512,0.029270221789677937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2048,1536,0.023744000328911677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,51200,1536,0.13466310501098633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,128,0.0031102223114834893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,64,0.0770311090681288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,4096,0.026555554734336004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,768,0.01092977739042706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,32,32,0.006205333189831839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,8192,0.02366133365366194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,2560,0.01238044434123569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2048,3072,0.03944355580541823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,51200,0.5009928809271919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,2560,0.008240000241332585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,6144,3072,0.053660445743136935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,256,0.0033857776886887024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,64,0.006478222293986215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,7168,0.0511448880036672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,256,0.004538666870858935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,32,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,65536,0.5639235708448622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,64,512,0.00387999994887246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,7168,0.07328533464007907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,4096,0.010848888920413123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,12288,0.17667110761006674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,32,0.013884444203641681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,51200,0.2258737881978353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,12288,0.01534400052494473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,256,0.00314666661951277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,1024,65536,0.7182177967495389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,12288,0.17008533742692736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,4096,0.019308444526460435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,768,1024,0.004243555582231945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2560,256,0.004743111216359668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1536,5120,0.011039111349317761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,1536,0.006021333237489064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,4096,10240,0.39131999015808105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,51200,2560,0.6438044442070855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,1536,0.06416355239020453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1024,10240,0.010812444819344414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,7168,0.08972266647550796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,128,4096,0.00958933350112703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,2048,0.0063973334100511335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,128,7168,0.01090133355723487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,7168,0.06084089146720039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,8192,0.02600266701645321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,65536,0.06812000274658203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,768,0.06892800331115723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,10240,0.026527111728986103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,12288,0.020989333589871723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,256,0.01955111159218682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,16384,0.16890045007069907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,768,0.028173334068722192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,51200,0.20415200127495658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,3072,1.2439244588216145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,6144,0.23938934008280435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,7168,0.11101688941319783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,768,0.02727288835578495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,1536,0.011912888950771756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3584,65536,1.8148061964246962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,2560,0.05119111140569051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,6144,0.02014400064945221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,16384,0.6056230862935384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,4096,0.34154844284057617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,16384,0.02683555583159129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,3072,0.0846782194243537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,1536,0.1277786625756158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,7168,0.015096889601813423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,12288,768,0.08487377564112346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,512,32,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,7168,0.024875554773542616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,256,0.003044444446762403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,1024,0.01128266668981976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,5120,0.013069333301650153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,512,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,7168,0.11238043838077122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,6144,2.1716605292426214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,4096,0.021344888541433547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,12288,8192,0.03616888986693488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,65536,12288,4.738753848605685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,1024,0.005848000033034219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,1536,0.013208000196350945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,512,0.018823999497625563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,8192,0.07872266901863946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,256,0.0052133335007561584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,32,0.01110666659143236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,2048,0.008559111091825698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,3072,1536,0.008586666650242275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,65536,7168,3.3362159729003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,4096,0.42488445176018613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,10240,256,0.006270222365856171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,3072,4.558233049180773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,65536,0.1350675556394789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,64,768,0.0069777775141927935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,3072,0.011749332977665795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,64,12288,0.008591111335489485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,6144,0.015512888630231222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,2048,0.08134844568040636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,10240,0.028332445356580947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,2560,0.19563289483388266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,8192,8192,0.03444266650411818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,65536,8192,2.839809841579861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,5120,0.008734222087595198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,768,0.006190222170617845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,64,65536,0.28431110911899143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,16384,0.04518933428658379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,6144,4096,0.029465778006447688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,51200,0.3717333210839166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,3584,0.014444443914625378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,128,0.003080888961752256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,16384,0.9658258226182727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,10240,32,0.006139555739031897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,51200,0.1381208896636963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,51200,0.04762844575775994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2560,1536,0.007135999699433644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,51200,5.919482760959202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,6144,2048,0.016537777251667447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,1024,0.08336799674563938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2048,64,0.006884444504976273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,2560,0.02034133341577318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,2560,1.645567152235243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,10240,0.12399021784464519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,2560,4096,0.013935110635227628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,7168,0.6507875654432509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,256,0.005868444426192178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,512,0.009224889179070791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,8192,2560,0.01961955593691932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3072,65536,0.08316889074113634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,256,0.009707555174827576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,6144,0.03810755411783854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,3072,0.015690666106012132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,768,6144,0.09689333041508992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,16384,2.6688844892713757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,5120,0.16537866327497694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,51200,0.5849484337700738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,32,0.14274133576287165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,7168,0.0388026667965783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,5120,1536,0.06751555866665311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,768,0.3399653434753418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,2048,0.04228622383541531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,512,256,0.005193777796294955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,10240,0.03818488783306546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,16384,0.07887644237942167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,10240,51200,1.5336035622490778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,256,1024,0.0075128889746136135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,128,768,0.005939555664857228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,51200,256,0.010827555424637265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,768,64,0.006897777732875612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,8192,0.01796711153454251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,2048,0.00831288927131229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,1024,512,0.011197333534558615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,6144,0.04976800084114075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1024,1536,0.0075626663035816616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,4096,0.008871999879678091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,51200,1.2024720509847004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,2560,0.011131555669837527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,128,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,8192,12288,0.6274684270222982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,12288,1.0461066563924153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,7168,0.03316266669167413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,768,0.015215999550289579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,7168,0.05466844307051765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1536,51200,0.0746871100531684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,2048,0.44822221332126194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,2048,0.14349867237938776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,5120,0.014514666464593677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,6144,4096,0.016024889217482675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,7168,65536,7.214725070529514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,10240,0.029327998558680218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,5120,0.15519377920362684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2560,65536,0.07274399863349067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,4096,64,0.013973333769374423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,128,0.006640888750553131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,256,4096,0.04033244318432278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,768,768,0.00490488898422983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,12288,0.016895999511082966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,5120,8192,0.03426933288574219
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,3072,0.3883768982357449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,3584,0.01127199994193183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,32,0.06111644373999702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,256,0.008553778131802877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,5120,0.17683822578854033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,1024,0.005420444326268301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,1536,0.023991111252042983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,12288,0.3374026616414388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,256,64,0.008576889004972246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1536,3072,0.05729689200719198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,7168,0.020465777979956735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,1024,0.010495999621020423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,1024,0.011195555329322815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,256,0.01734844512409634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2048,768,0.006335999816656113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,128,256,0.005577777822812398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,12288,0.5033493571811253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,16384,0.039466665850745306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,768,32,0.005462222215202119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,512,0.015206222732861837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,1024,0.2880684534708659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,8192,65536,0.19869600401984322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,6144,3584,0.01583377851380242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,8192,0.011780444118711682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,768,0.006114666660626729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,12288,0.0064444442590077715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,128,0.005369777894682354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1536,5120,0.014363555444611443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,8192,0.047600001096725464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,51200,6.587755415174697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,51200,1.3688444561428492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,65536,5120,1.8714737362331813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,256,0.006620444357395172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,4096,0.09226222170723809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,8192,12288,0.5746969117058648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,768,0.01072177787621816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,1024,0.011576889289749993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,768,0.004701333327425851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,12288,0.02327288852797614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,10240,0.0652551121181912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,16384,0.009452444811662039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,3072,0.14253155390421549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3584,512,0.0038968887594011095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,1024,0.03790222273932563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,32,1024,0.004022222426202563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1536,65536,0.6381493144565159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,16384,2048,0.21137244171566436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,2048,0.010651555326249866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,12288,0.09102755784988403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,2560,0.007527111305130853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,32,64,0.0074728892909155945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,8192,0.04617066515816582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,1024,0.0053688887920644544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,1024,0.004043555508057277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,8192,0.0665280024210612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,10240,0.01907822158601549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,128,0.0053226666318045715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,4096,0.026260443859630164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,51200,16384,0.4262773460812039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,256,0.008054222497675154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,768,0.006856888946559694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,5120,0.0096133334769143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,3584,0.020017777880032856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,256,0.005530666559934616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,3584,0.013076444466908773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,12288,0.047154668304655284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,2048,0.007081777685218387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,256,0.0037564444873068067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,8192,0.15034843815697566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,4096,0.07367822196748522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,128,0.004023999803596073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,1024,0.033321777979532875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,3072,2560,0.42151912053426105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,3584,0.11824977397918701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,2560,0.01039644413524204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,512,16384,0.06751022073957655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,12288,0.10431733396318223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,10240,1024,0.013632888595263163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,256,0.019919999771647982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,2560,0.014626666903495789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,1024,0.015184000134468079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,512,0.006511999915043513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,7168,0.03141422073046366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,2560,0.019462222854296368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,3072,512,0.004957333207130432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,16384,0.1365600029627482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,65536,512,0.21316177315182158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,2560,4096,0.07963199747933282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,65536,0.07346844673156738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,256,3584,0.024991111622916326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,64,16384,0.018650665879249573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,4096,0.011271110839313932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,1536,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,1536,0.016102222932709586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,4096,0.017384888397322763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,256,0.0069617778062820435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3584,5120,0.012294222083356647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,6144,0.006261333409282897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,12288,8192,1.104699558681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,51200,0.4040986696879069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,1536,0.00867555538813273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,2560,0.007585778004593319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,1024,0.026736888620588515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,16384,128,0.025290666355027094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,12288,2560,0.3422444396548801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,5120,0.0936711099412706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,8192,8192,2.0221741994222007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,128,0.005535111245181825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,512,0.007098666495747036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,2048,0.01733955575360192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,1024,0.007043555378913879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,3584,64,0.014908444550302295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,3072,0.3019351164499919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,512,8192,0.008622222476535374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,3072,0.011685333318180509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2048,8192,0.07154488563537598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,51200,0.060714668697781034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,6144,0.10036533408694798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,6144,6144,0.03569688730769687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,16384,0.019893333315849304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,65536,0.2614719867706299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,3072,0.056201775868733726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,2560,0.007168889045715332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,256,3072,0.0069271110826068455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,3584,0.020063999626371596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,32,0.0038533334930737815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,4096,256,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,16384,0.4220302104949951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,10240,1.3145217895507812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,64,0.0035128887328836652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,2560,0.054973334074020386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,10240,0.01941066649225023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,1024,0.007090667055712805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,1536,0.0073928890956772705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,6144,1.4550364812215169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,7168,0.9647608862982856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2048,7168,0.015234667393896313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,12288,0.1499075492223104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,1024,0.043493333790037364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,64,65536,0.00925066653225157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,10240,0.6890835762023926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,6144,0.2786124547322591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,1024,0.006609777609507243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,128,7168,0.09932000107235378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,64,2048,0.007000000112586551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,2560,8192,0.026087111896938745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,12288,0.8035928938123914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,512,0.0654275549782647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,10240,0.03968088825543722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,256,0.15815911028120253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,32,0.0032231110251612137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,768,0.014545778433481852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,1024,0.010460444622569613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,128,0.004262222184075249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,51200,128,0.016399999459584553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,51200,0.2692631085713704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,5120,0.0791777769724528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,32,0.005225777626037598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,64,256,0.005249777601824866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,768,0.010480889015727572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,32,768,0.005727111051479976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,2560,0.013033777475357056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,768,0.24987022082010904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,3072,0.019426667028003268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,1536,0.005505777895450592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,12288,6144,0.04111288984616598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1536,6144,0.015704888436529372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,16384,65536,0.29936710993448895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,3072,0.006447111152940326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,16384,51200,7.359830220540364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,51200,0.0888062251938714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,10240,2048,0.016511999898486666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,512,0.014902222487661572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,51200,0.07740177710851033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,2048,0.7249404589335123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,128,0.004042666819360522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,6144,0.07474399937523736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,768,0.0058124446206622655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,128,32,0.002500444443689452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,32,0.028743998871909246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,10240,0.33782222535875106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,256,2560,0.018949333164427016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1024,256,0.0052639999323421055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,5120,0.035255110926098294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,6144,2048,0.010107555323176915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,4096,0.03527733352449205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,128,0.0041440000964535605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,5120,0.1653226613998413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,65536,0.2756560113694933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,51200,0.9732924567328559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,10240,1536,0.01404711107412974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,7168,0.014766222900814481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,3072,0.013439999686347114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,8192,768,0.08024888568454318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,8192,0.05100088980462816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,6144,0.029956443442238703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,7168,0.114830215771993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,1024,0.004311111238267687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,4096,3584,0.0214684440029992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,5120,0.04575644599066841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,10240,12288,0.045407109790378146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,8192,6144,0.13133777512444392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,1536,0.00685066646999783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1024,512,0.005127999931573868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,6144,0.04706222150060865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,65536,0.07938400242063734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,1024,0.0069448889957533936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,512,0.021935110290845234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,4096,0.017198221551047433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,1536,8192,0.011773332953453064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,32,0.003075555587808291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,7168,0.06338755289713542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,8192,0.06193777587678698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,4096,0.05921155876583523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3072,256,0.005423999908897612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,10240,0.019172444939613342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,65536,5120,1.4343751271565754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,5120,0.036882665422227644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,6144,65536,2.4595599704318576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,2560,0.013190222283204397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,32,0.008015111088752747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,3072,0.022312000393867493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1536,10240,0.19135022163391113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,512,0.0034764446318149567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,5120,128,0.15080000294579401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,10240,0.015376889043384127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,12288,32,0.007887110941939885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,5120,128,0.004770666774776247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,2560,0.007582222421964009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,128,0.004848000076082018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,1536,0.006079111248254776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,512,0.005944889038801193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,512,768,0.012974222501118978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,3072,0.009757333331637913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,2560,0.00777333312564426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,8192,0.01184977756606208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,4096,64,0.021576889687114294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,51200,0.06120622158050537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,6144,3584,0.1675342188941108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,2048,0.019998222589492798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,1024,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,7168,128,0.004400889078776042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,64,0.005386666705211003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,8192,0.09754400120841132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,3072,0.008995555341243744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,1024,0.027628444963031348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,7168,0.03771377934349908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,65536,10240,2.6912356482611766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1536,65536,0.09531378083758885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,64,0.017312889297803242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,512,0.060469335979885526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,7168,0.06186133623123169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,7168,12288,0.04680800106790331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,256,0.004365333252482944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,3584,5120,0.12339288658565944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,16384,0.04548266530036926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,768,7168,0.06641688611772326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,3584,0.22088355488247344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,3584,0.1934800015555488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,64,0.005474666754404704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,128,0.006136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,16384,1536,0.013726222018400827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,2560,2048,0.006968888971540663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,65536,512,0.10494933525721233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,256,16384,0.01943555474281311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,10240,0.22715732786390516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,768,2048,0.012660444610648684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,6144,0.028041776683595445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,7168,7168,0.20184177822536895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,512,0.005103111267089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,2048,0.08592089017232259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,2560,0.007128000259399414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,4096,1024,0.005767111149099138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1024,128,0.004967999955018361
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,128,0.002871111035346985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,8192,0.06372621986601087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,4096,0.006230222268237009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,4096,8192,0.0262097782558865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,65536,0.06952977842754789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,64,0.0027075554761621687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,768,256,0.0042568888101312844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,10240,3072,3.163252512613932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2048,1536,0.00868800034125646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,1536,0.04645333356327481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,3072,0.2245173454284668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,128,65536,0.03540800015131632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,256,0.007368889119890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,8192,0.0092622223827574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,10240,0.018191110756662156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,7168,0.010928000013033548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,4096,0.011352000137170156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,32,64,0.002584888910253843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,6144,0.010997333460383944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,8192,6144,0.13743200567033556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,1024,0.009388444324334463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,4096,3072,0.011183111204041375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,65536,4096,0.07271822293599446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,2560,0.04618666569391886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,65536,1.1453350914849174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,3072,0.008088889221350351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,1536,0.007711110843552484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,4096,0.19955289363861084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,1024,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,512,0.004074666649103165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,7168,0.02362666692998674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,1024,16384,0.751164436340332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,512,0.005687110953860813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,5120,1536,0.008169777691364288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,5120,3072,0.11227467324998643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,3584,0.07516888777414958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,256,0.003321777822242843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,256,4096,0.010130667024188572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,1536,0.07231199741363525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,65536,2048,0.12451910972595215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,65536,2560,0.04810488886303372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,64,0.0058062221441004015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,512,0.021009777983029682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,7168,0.015408888459205627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,10240,0.28042133649190265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,16384,0.20474399460686576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,3584,0.008894222478071848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,1536,0.010226666927337646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,256,5120,0.0068400001360310455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,51200,0.06488444407780965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,3584,0.008943110704421997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,5120,8192,0.22424533632066515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,12288,1536,0.012854222622182635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,3072,0.0163111107216941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,256,0.029046220911873713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,768,256,0.003526222374704149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,10240,3072,0.016336000627941556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,768,0.007143999967310164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,256,0.003802666647566689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,16384,0.031094223260879517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,5120,0.0807360013326009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,1536,0.007271111011505127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,51200,3584,0.9255751503838433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,1024,0.0058471109304163195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,2560,1.1896968417697482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,51200,5.952240837944879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,51200,1024,0.024307555622524683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,8192,0.00850933293501536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,128,0.023376888699001733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,1536,0.007272000114123027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,2048,0.007251555720965068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,12288,0.006634666687912411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1024,2048,0.007344000041484833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,7168,0.10444888803693984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,64,0.003383999897374047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,256,2560,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,768,2560,0.007993777592976889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3072,2560,0.01053244454993142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2560,51200,0.08369600110583836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2048,12288,0.0229768885506524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,512,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,256,0.004871110949251386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,51200,128,0.006719111154476802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,64,0.007452444069915348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,51200,0.05561511384116279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,5120,0.03678844372431437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,256,0.00573511090543535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3584,256,0.014640889234013028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,8192,3584,0.013541333377361298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3584,1536,0.006493333313200209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,32,0.005005333158704969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,5120,4096,0.01829777823554145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,16384,0.12407555845048691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,4096,0.00795111142926746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,3072,128,0.004257777912749184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,768,0.024383111132515803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,51200,0.051345778836144336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,16384,2048,0.1456035508049859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,768,0.290063116285536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,256,0.029792000850041706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,12288,12288,0.9056648678249783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,3072,0.017441777719391715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,768,0.009484444227483537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,256,2048,0.005389333185222413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,7168,0.028757333755493164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,3072,1536,0.032998220788107976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,512,256,0.00300799993177255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,3584,0.008007110820876228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,5120,0.09871555699242486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,51200,0.0321413344807095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,7168,0.042620443635516696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,4096,64,0.006486222147941589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1024,65536,0.0861884421772427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,1536,0.04732799861166212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,2048,0.06587555673387316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,51200,0.3326968881818983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,3072,0.007996444072988298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,3584,0.015471999843915304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,3072,0.007224889265166388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,64,5120,0.01036622209681405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,1024,0.012711111042234631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,7168,2560,0.11791111363304986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,3584,0.0682835578918457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,16384,0.024158222807778254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,5120,0.08007199896706475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,2560,0.006883555402358373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,3584,0.03856888744566176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,256,7168,0.02471911079353756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,3072,0.014001778430408902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,2048,0.011179555621412067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,128,1024,0.004025777594910728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,7168,0.03548177745607164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,65536,256,0.019910222954220243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,10240,0.36363644070095485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,128,0.006624889042642381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,6144,1.615043534172906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,7168,0.02903288933965895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,6144,0.025090666280852422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3072,32,0.015600000818570455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,4096,0.010977778169843884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,1536,0.012587555580668978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,768,0.01423822177780999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1024,2048,0.013552000125249227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,1024,12288,0.022738666998015508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,1024,0.038932445976469256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,6144,16384,0.09426222244898479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,32,0.004475555486149258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,256,0.01072799993885888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,12288,0.07847111092673408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,128,0.005105777747101254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,10240,0.6112426651848687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,65536,0.07807466718885633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,1024,0.07594399982028537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,7168,0.10817955599890815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,2048,0.21717688772413465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,6144,0.01404800017674764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,512,0.062336001131269664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,32,2560,0.005642666584915585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,2048,0.3970906734466553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,8192,0.043623109658559166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,65536,5120,33.77851698133681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,2560,0.0221413340833452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2560,3584,0.07514666848712497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,7168,0.053118222289615206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2560,4096,0.011947555674446953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,6144,768,0.026695999834272597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,4096,16384,0.2210977739757962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,6144,0.011572444604502784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,8192,0.27602044741312665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,256,2048,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,512,5120,0.05425066749254862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,3072,0.03199911117553711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,7168,0.019179556104871962
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,12288,0.017543110582563613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,4096,0.42116088337368435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,64,0.004313777718279096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2048,128,0.009893333746327294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,128,0.024975111087163288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,3072,0.012687111066447364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,10240,0.046000887950261436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,256,0.005843555761708154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2560,128,0.014915555715560913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,1536,0.040287110540601946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,3072,0.010077333284748925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,512,0.20000089539421928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,65536,0.6149253315395778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,8192,2048,0.7887022230360242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,64,0.003585777762863371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,12288,6144,0.04182755615976122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,16384,0.22576000955369738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,512,0.01702577703528934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,64,8192,0.016689777374267578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,1536,12288,0.01717333330048455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2048,2560,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,5120,2560,0.08377599716186523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,1024,0.02151022189193302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,6144,4096,0.020000000794728596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3072,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,128,0.005015999906592899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,512,0.025913778278562758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,65536,0.2898302343156603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,4096,0.10535644160376655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,128,0.024826667375034753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,8192,0.02016177773475647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,51200,2048,0.4253031147850885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,32,0.005320888840489917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,5120,0.5073217815823025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,5120,0.016923555069499545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,12288,0.5754568841722276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,3072,0.14511111047532824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,7168,1024,0.007491555478837755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,65536,4096,0.1344053347905477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,1536,0.018790221876568265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,51200,0.5714613596598307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,1536,0.014504889647165934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,1536,0.10431289010577732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,128,7168,0.010574222438865237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,51200,2048,1.1333573659261067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,12288,6144,0.04131822122467889
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,512,12288,0.022351110974947613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,512,0.0069857777820693115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2048,16384,0.024520888924598694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,128,0.006207110981146495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,4096,0.010599999792046018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,6144,0.04579200016127693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,5120,256,0.003996444245179494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,3072,0.008007110820876228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3072,8192,0.021328000558747187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,256,16384,0.007924444145626491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,768,12288,0.02018222212791443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,1536,0.023172444767422144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1024,51200,0.6306791305541992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,3584,0.013193777865833707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3072,256,0.004744889007674323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,768,0.04025955663786994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3072,512,0.003579555700222651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,128,512,0.0053119998839166425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,32,0.006716444260544247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1536,10240,0.017852443787786696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,128,0.0028479999552170434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,2560,0.2341991000705295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,32,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,256,0.012554667062229581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2560,8192,0.10811822281943427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,10240,0.009574222067991892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,5120,12288,0.36951200167338055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,51200,0.06222933530807495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2560,64,0.017921777235137094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,1536,0.035183111826578774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,128,768,0.009102221992280748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,256,0.003026666740576426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,6144,4096,0.01661600006951226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,65536,0.9046630859375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,5120,0.014144889182514615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,2048,0.10959555705388387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,2560,0.08188622527652316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,128,0.0027662221756246355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,768,0.020996444755130347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,51200,0.11634577645195855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,512,1536,0.011154666543006897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,10240,256,0.03754222061898973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,4096,768,0.006786666810512543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,6144,0.013026666310098437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,65536,0.030228445927302044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,128,0.010857777463065254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,512,0.016208888755904306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,2048,0.006574222197135289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3072,2560,0.051185776789983116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,64,0.007828444242477417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,256,0.018292443619834054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,6144,16384,0.051575111018286816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,512,0.011575111084514193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,7168,65536,1.9860907660590277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,2048,0.016136000553766888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,3584,0.0932604471842448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,128,4096,0.020568889048364427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,16384,0.0498906672000885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,6144,0.010803555448849997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,1024,0.007075555622577667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,4096,0.03264800045225356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,65536,0.2542240089840359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,8192,1024,0.00943199959066179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,64,768,0.00582755563987626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,65536,0.06804800033569336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,768,1536,0.005018666800525454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,65536,4.928131527370877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,2560,0.030618667602539062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1536,2560,0.009680888719028896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,8192,768,0.06124355395634969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,256,0.14023021856943765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,4096,0.061012446880340576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,8192,0.016040000650617812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,6144,16384,0.06070044305589464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,1024,0.006659555352396435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,4096,0.014676445060306124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,12288,0.5425688955518935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,32,65536,0.04476622078153822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,4096,0.3215555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,1024,0.004984000076850255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,3072,0.033815112378862165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,128,51200,0.24729779031541613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,512,0.006814222368929122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,128,12288,0.5692630873786079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,3584,0.007790221936172909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1024,5120,0.2192088895373874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,4096,0.2123119963539971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,768,0.0038693332009845306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,12288,0.01311822235584259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,5120,3584,1.7649493747287328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,1536,0.0064382221963670515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,3072,0.48089509540134007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,12288,4096,0.022672888305452134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,3584,0.018927110566033255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,7168,0.14498399363623724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,10240,0.8643369144863553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,2560,0.5500337812635634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,8192,0.018612444400787354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,8192,0.04918044474389818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,1024,0.005585777676767773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,8192,65536,0.30122489399380153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,512,1024,0.006147555592987273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2048,512,0.006424888968467712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,1024,0.007888889147175683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,32,0.006151999864313338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,16384,0.4089137713114421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,1024,0.05354933275116814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,8192,0.26298756069607204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,2560,0.14187822077009413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,10240,0.024254222710927326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,10240,16384,0.7754853566487631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,1536,0.035465776920318604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,1024,0.009082666701740688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,7168,0.784711996714274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,4096,1536,0.32753245035807294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,512,0.0032266666077905228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2560,10240,0.3266399966345893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,32,6144,0.04669955703947279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,1024,0.006574222197135289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,64,12288,0.011901333100265928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,512,3072,0.008992888861232335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,8192,0.06554666492674086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,512,0.03312177790535821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,1024,4096,0.010961777634090848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,12288,0.16859200265672472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,16384,1.5058711369832356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,51200,0.013825777504179211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1536,4096,0.009415999882751042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,8192,0.07211822271347046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,51200,2.6048319074842667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,2048,0.013781333135233985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,128,512,0.007419555551475949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,10240,16384,0.8972319497002496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,8192,0.21020978026919893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,7168,0.019398222366968792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2048,128,0.007265778051482306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,8192,65536,3.700170728895399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,1024,0.0060159998635451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,768,0.006353777729802662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,128,0.005689777847793367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,51200,8192,0.20025600327385795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,6144,0.10213955243428548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,16384,1.2771067089504666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,6144,0.049606220589743726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1536,2560,0.03911644551489089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,128,0.009826666778988307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2560,768,0.0058311112225055695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,3584,0.008316444853941599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,2560,0.007135999699433644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,64,0.00360622215602133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,7168,2048,0.08999466896057129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,5120,0.01294133315483729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3584,512,0.02388711108101739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,128,0.0032888888898822996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,6144,0.5324213239881728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1536,768,0.005079111291302575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,5120,0.02590399980545044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,51200,0.20136800077226427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,512,0.022501332892311945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,8192,0.18768178092108834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,3072,0.07102666960822211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,16384,3584,0.04659733176231384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,8192,51200,1.562429322136773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,7168,0.18839200337727866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,32,0.006594666590293248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,768,0.007050666544172499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,5120,0.10083911154005264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,3072,0.00959822204377916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,2560,0.008536000218656328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,3584,0.6024808883666992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,3584,4096,0.04477511180771721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,4096,0.02681422233581543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,32,0.026509332987997267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,6144,6144,0.17774577935536703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,128,0.01129066695769628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,64,0.0027644443843099806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,5120,4096,0.12675821781158447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,12288,0.07882222202089098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,7168,16384,0.6695111062791613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,768,0.026132444540659588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,5120,0.10084266795052423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,51200,0.20349155531989205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1536,2560,0.00944800012641483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,65536,0.07545599672529432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,7168,0.029696888393825952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,128,7168,0.008245333201355403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,6144,8192,0.029141333368089464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,5120,0.016366221838527255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,3072,0.08477333519193862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,3072,0.007189333438873291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,5120,0.010323555933104621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,256,0.0029111111329661477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,768,0.0034755555291970572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,3072,0.006240889016124938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,5120,0.30209244622124565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,51200,1.8384079403347438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,256,10240,0.026662222213215295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,51200,0.31196710798475474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,12288,0.06431021955278185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,5120,8192,0.1564400063620673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,51200,0.19641333156161836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,128,64,0.005158222383923001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,3072,0.21178578005896675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,3072,0.0787608888414171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,7168,5120,0.015574221809705099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,6144,0.20954844686720106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,128,0.003031111011902491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,12288,0.02944533361328973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,12288,0.013418667018413544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,10240,0.016368000043763053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,1536,0.07607377900017633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,2048,0.5421991348266602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,5120,0.012759110993809171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,5120,3584,0.09888977474636501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,51200,128,0.0990862250328064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,768,0.010484444598356882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,128,0.0029208889851967492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,128,6144,0.024118221468395654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,51200,0.133443554242452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,1536,0.35253243976169163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2560,2560,0.03164533442921109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,65536,16384,0.2823066711425781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,32,0.0027502222607533136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3072,768,0.01824355622132619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,8192,0.12350755267673069
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,16384,512,0.11177866988711888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,8192,3072,0.014096889230940076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,12288,0.6048311127556695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,5120,0.008451555338170793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,16384,0.07823999722798665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,16384,0.746687094370524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,6144,0.019496889577971566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,5120,0.09229689174228245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,12288,4096,0.020399999287393358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,3072,0.007463999920421177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,128,32,0.00655733338660664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,2560,256,0.005590222362014983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,3072,0.03048711021741231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,768,16384,0.20066489113701713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,2048,0.013463111387358772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,16384,1536,0.013955555028385587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,64,0.005134222408135732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,512,0.005315555466545953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,32,512,0.005497777627574072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,2048,0.04369866516855028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,1024,16384,2.1207608116997614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,5120,0.01035644445154402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,64,2560,0.007671111159854465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,512,0.0033973333322339584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2560,32,0.005138666679461797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,7168,1.0046621958414714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,5120,0.03407111101680332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,256,0.005659555395444234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,8192,0.0788853300942315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,3584,0.06348799996905856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,6144,0.012800888882742988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,4096,0.014648889501889547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,16384,3072,0.019332443674405415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,128,0.004419555680619346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,512,128,0.005559999909665849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,16384,0.2928080028957791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,3584,4096,0.0732462207476298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,3584,0.00942755573325687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2048,3584,0.043996443351109825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,3072,0.017824000782436795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,16384,0.00963288876745436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3072,10240,0.01863911085658603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,512,5120,0.5584826469421387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,256,0.0032888888898822996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,768,0.025324443976084392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,256,8192,0.42282666100396055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,128,0.003937777959638172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,12288,0.027191110783153113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,128,0.007536888950400882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,51200,2.234018749660916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,12288,0.045997334851158984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,2048,0.06580888562732272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,3072,0.024338665935728285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2048,12288,0.1238480011622111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,5120,0.059508442878723145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,3584,0.006781333436568578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,3584,0.13144000371297201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,768,0.020384889509942796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,32,0.0025457777082920074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2048,2560,0.0830355551507738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,5120,0.0614195532268948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,65536,0.9163413577609592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,3072,0.053005334403779775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,512,128,0.0034346667428811393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,12288,12288,0.07922310961617364
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,65536,0.768580436706543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,51200,8192,0.5604204601711696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,32,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,2560,0.011598221957683563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,10240,0.01847822301917606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,128,3584,0.02590844366285536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,256,0.005502222312821283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,64,0.0058000000814596815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,32,0.0052284445199701525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,6144,0.022797332869635686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,64,0.0052168890833854675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,128,32,0.0034844444857703317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,768,0.006511111226346757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,16384,0.11019911368687947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,6144,0.1983013285530938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,3072,0.0274000002278222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,768,0.0063582224150498705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,6144,0.014471999473041959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,6144,64,0.00370755543311437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,2048,0.05192977852291531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,2560,0.25508801142374676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,1536,12288,0.026320000489552815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2560,64,0.0034355556385384668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,8192,0.07078488667805989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,1024,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,12288,768,0.08556355370415582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,1536,0.013616888887352414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,51200,12288,4.896305931939019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,8192,128,0.018704889549149405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3584,128,0.010614222122563256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1536,2048,0.03390311201413473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,10240,3072,0.08726044495900472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,1024,768,0.005222222043408289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,6144,5120,0.01833599971400367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,2560,32,0.002793777734041214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1024,1024,0.005807999935415056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,10240,0.014209777116775513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,128,0.006612444503439798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,2560,1.5476399527655706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,4096,0.21147110727098253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,8192,1536,0.0811413327852885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,8192,0.023200000325838726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,1536,0.012960000170601739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,6144,0.009505777723259395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,3072,0.18524089124467638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,2560,768,0.00555288874440723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,16384,0.026944888962639704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,5120,0.13166577286190456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,5120,128,0.005489777773618698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,16384,16384,0.25358666314019096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,2560,7168,0.19676444265577528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,32,1536,0.00434311106801033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,65536,0.29919645521375865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,3584,0.11335910691155328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,4096,0.008039111064540016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,10240,0.08365599976645575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,768,0.02051911089155409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1024,51200,0.05275911092758179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,65536,15.61265394422743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,256,1024,0.008593777815500895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,1024,0.011821333732869891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,8192,0.25815823343065053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,8192,0.01953866746690538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,12288,0.015136000182893542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,10240,4096,0.026314665873845417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,2048,0.01977333260907067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,6144,0.012282667060693106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,16384,10240,0.30959200859069824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,3584,0.03204710947142707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,10240,0.25563110245598686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,7168,0.02546311087078518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,512,6144,0.029680000411139593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,3072,0.008749333520730337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,1024,1536,0.022615111536449854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,8192,0.04226933254135979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,2560,0.011295999917719098
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,768,0.004176889028814104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,65536,12288,0.553400887383355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,1024,3584,0.04962311188379923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2560,7168,0.02738488879468706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,65536,0.12930222352345785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,65536,0.18789422512054443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,4096,3072,0.012138666378127204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,256,0.004869333157936732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,2560,3072,0.019048889478047688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,512,5120,0.02572444412443373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1536,256,0.0033039999090962936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,4096,2560,0.039275556802749634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3072,4096,0.06226311127344767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,64,6144,0.005888888819350137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3584,12288,0.02697155541843838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,512,0.005550222264395819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,2048,4096,0.01069155583779017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,1024,0.004921777794758479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,51200,0.07982399728563097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,1024,0.005910222315125995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,6144,768,0.024307555622524683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,768,0.006593777901596493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1024,3584,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,4096,0.3612728913625081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,65536,0.41825244161817765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,1024,0.024248000648286607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,128,6144,0.0665831102265252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,32,10240,0.013388444152143268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,3584,0.006524444454246097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,4096,768,0.005464000006516774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,1024,0.00600622221827507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,64,0.035593777894973755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,6144,128,0.018386666973431904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3072,8192,0.014256000518798828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,10240,0.05659821960661146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,8192,0.04101777738995022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,12288,256,0.006059555543793573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1536,128,0.010077333284748925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,12288,0.045885334412256874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,1024,0.004927111168702443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,8192,3584,0.01595288846227858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,16384,1.2204230626424153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,32,512,0.006674666785531574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,7168,0.014363555444611443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,12288,0.13579289118448892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,1536,0.012423999607563019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,7168,768,0.018087999688254464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,12288,6144,0.06326399909125434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,65536,8192,0.2834435568915473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,4096,51200,0.06976266702016194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,1024,0.009358222285906473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,6144,2048,0.07879200246598986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,3584,0.00794666674402025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,7168,0.1988959974712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,768,7168,0.03578044308556451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,128,0.0064151109092765385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,10240,7168,0.5933306482103136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3072,4096,0.11832088894314235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,12288,65536,0.2573964330885145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,8192,0.015012444721327888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,768,0.005953777581453323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,32,0.0057857777509424425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3072,8192,0.2224106788635254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,3072,0.46799643834431964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,512,0.021664889322386846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,4096,0.14521333906385633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,128,0.003724444243643019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,256,0.014140443669425117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,2560,0.030543999539481267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,768,0.024883555041419134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,12288,4096,0.244466675652398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,65536,0.10771289136674668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,16384,1.9574809604220922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,256,1024,0.005683555371231503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,2048,5120,0.016401777664820354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,12288,0.02358666724628872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,7168,0.18599288993411592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,2560,0.021888888544506494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2560,2048,0.009016888837019602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,128,4096,0.013487111363146039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,4096,0.14282488822937012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,128,0.006194666855865055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1024,1536,0.00833422193924586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,4096,0.052988443109724254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,2560,0.013325333595275879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,1024,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,10240,0.014120888378885059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,65536,0.16094844871097141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,10240,8192,0.04173422190878126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,12288,256,0.007788444558779399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,1536,0.018615111708641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,256,7168,0.03257511059443156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,4096,0.01907911068863339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,12288,2.824955622355143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,4096,0.013058666553762225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,6144,12288,0.6220169067382812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,32,768,0.0035573335157500375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,768,16384,0.037928889195124306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,5120,0.036632001399993896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,10240,64,0.04411733481619093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,7168,0.6851386494106717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,12288,0.5901110967000326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,768,0.005011555635266834
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,64,0.01677688956260681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,2048,0.020676443974177044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,128,0.005205333232879639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,5120,0.04287733303176033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1024,3584,0.031161778502994116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,65536,0.533273802863227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,7168,7168,0.15245866775512695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,2560,0.011709333293967776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,1536,51200,0.0770933363172743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,7168,16384,1.4865466223822699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,51200,1.6196293301052516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,10240,0.013306666579511432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,128,2560,0.010448888772063784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2048,4096,0.01649599936273363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2048,5120,0.02722666660944621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,10240,65536,21.75717502170139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,10240,3072,0.03481600019666884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,6144,0.03387644555833604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,2560,0.010313777459992303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,51200,0.42188178168402773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,2560,0.05052266518274943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,12288,0.02151022189193302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,128,0.005727111051479976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,768,0.007634667058785756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,12288,2560,0.02641422218746609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,65536,0.49115822050306535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,32,0.16008978419833714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,12288,0.024079999989933435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,3072,0.025416889124446448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,51200,0.24030311902364096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,3584,0.9012141757541232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,1536,0.07173244158426921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,2048,0.021312889125612047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,7168,65536,0.14119111167060003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,1024,0.080494225025177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,1024,0.06391200092103747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,16384,1.0049875047471788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,16384,1536,0.015107555521859063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,2560,0.008214222060309516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,3072,0.10198044114642674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,10240,0.02371999952528212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,16384,0.12068622642093235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,5120,256,0.02602400051222907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,256,0.19695466094546846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,6144,0.17924977673424614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,256,0.003839111162556542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,256,0.007027555671003129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,51200,0.26968267228868276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,768,1536,0.006384000182151794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,7168,0.011115555961926779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,1024,2560,0.00867733359336853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,65536,5.502246432834202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,768,16384,0.17519377337561715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,7168,0.036582221587498985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,10240,3072,0.257135099834866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,128,0.00904444439543618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,4096,8192,0.016594666573736403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,65536,1024,0.2021324502097236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3584,16384,0.05829155445098877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,16384,51200,0.4226871066623264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,16384,0.03802400165134006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,3584,0.01643200053109063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,4096,0.03734222385618422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,10240,0.33191466331481934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,768,0.010464000205198923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,64,0.008153777983453538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,3584,0.06625511248906453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,5120,12288,0.035162667433420815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,256,0.14614400598737928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,512,0.0060159998635451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,512,0.005340444544951121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,65536,0.3862906561957465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,6144,0.008115555677149031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,5120,64,0.0033128888656695685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,8192,0.11135644382900661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,256,0.005232000102599462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,2560,0.012798222402731577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,768,0.004675555560323927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,128,0.0075902218619982404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,4096,1536,0.016584888100624084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,12288,0.02993866801261902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3072,6144,0.012898666991127862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,10240,6144,0.657808886633979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,12288,0.03452533483505249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,4096,1536,0.05316977699597677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,12288,0.4130515522427029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,6144,12288,0.32386933432685006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,12288,51200,4.985303243001302
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,1024,0.16427377859751383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,51200,8192,2.414140489366319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,2048,0.47496265835232204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,128,0.010025777750545079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,128,512,0.017488888568348356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,7168,512,0.02369333306948344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,3584,0.4296862284342448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,1024,0.3683786657121446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,6144,12288,0.0745697749985589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,16384,0.019875556230545044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,12288,0.14205422666337755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,32,3584,0.008319999608728621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3584,256,0.005629333357016246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,256,0.0047564444442590075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,65536,64,0.00833777752187517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1024,512,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,64,0.015030221806632148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,8192,0.01611199975013733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,3072,0.009556444154845344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,6144,512,0.008162666526105668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,6144,4096,0.01843377782238854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,12288,0.012865777644846173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,768,0.006280888699822956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,6144,0.026779555612140234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,5120,0.016235555211702984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,1024,0.040836443503697716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,7168,0.280985779232449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,5120,6144,0.14436266157362196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,16384,1536,1.4511822594536676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,512,0.00703022215101454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,65536,3.646142323811849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,2048,512,0.015210666590266757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,16384,0.18115022447374132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,128,0.014548444085650973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,65536,0.11675466431511773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,64,512,0.008615111311276754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,16384,128,0.030949334303538006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,32,0.01258933295806249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,768,5120,0.09165866507424249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,64,0.005409777578380372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,12288,0.20828889475928414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,2048,0.04279288980695936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,512,2560,0.014578666951921252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,8192,0.0284222231970893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,1536,0.027416000763575237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,3072,16384,0.3469715648227268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,64,0.004376888689067629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,128,0.0029786665820413162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,16384,2048,0.2614275614420573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,7168,3584,0.04111466805140177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,16384,0.2605146567026774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,16384,0.5349413553873698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,1536,65536,0.23103732532925078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3584,512,0.022628444764349196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,512,0.005062222066852781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,32,128,0.006907555378145642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,7168,2560,0.010753778119881948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,4096,65536,0.1448382271660699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,3072,0.12691556082831487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,16384,0.014162666267818876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,128,0.004960000101062986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,65536,0.06576177808973524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,6144,0.12682310740152994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,7168,3072,0.01389600005414751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,3584,3072,0.011253332926167382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,2560,128,0.004640000147951974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,12288,0.3952799903021918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,768,0.005059555586841372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,16384,51200,24.07315741644965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,32,3072,0.01033688916100396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,1024,1.3124898274739583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,512,7168,0.013861333330472311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,16384,0.20636267132229277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,4096,0.022031999296612207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,7168,0.018011555075645447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,5120,1.1963360044691298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,6144,12288,0.06695377826690674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,128,0.004943111290534337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,512,0.00738666703303655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,128,5120,0.011582222249772815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,4096,6144,0.14219021797180176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,7168,65536,1.8856879340277777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,8192,1536,0.020120888948440552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,512,0.005109333329730564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,128,0.005034666508436203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,64,256,0.005667555663320754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,128,6144,0.012852444416946836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,32,0.01055466632048289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,6144,10240,0.02817866537306044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,768,1024,0.0075146663520071245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,8192,0.02537511123551263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,64,4096,0.01126488877667321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,7168,128,0.006372444331645966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,8192,1536,0.00980355590581894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,16384,10240,0.8619324366251627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,32,7168,0.0384897788365682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,768,6144,0.013598221871587964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,512,0.013807999591032663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,256,8192,0.060384001996782094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,8192,512,0.0069351109365622205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,2560,0.007698666718271043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,2560,0.008829333715968663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2048,64,0.004994666824738185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,256,1536,0.00480000012450748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,128,0.00699733363257514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1024,65536,0.948287116156684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,3584,0.012324444121784635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,2048,128,0.011821333732869891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,256,0.013020444247457715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,8192,0.013931555880440606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,51200,0.11268177297380234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,64,0.04194666610823738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,768,65536,0.7402613427903918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,8192,1.8434293535020616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,51200,2560,0.3836924499935574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,128,0.14206222693125406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,256,65536,0.015230221880806817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,5120,0.026947556270493403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,6144,0.16733511288960776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,6144,5120,0.15478132830725774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,32,16384,0.006567111031876669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,256,8192,0.03342311249838935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,512,16384,0.10070933236016168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2048,6144,0.14921866522894964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,128,10240,0.01670222149954902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,128,3072,0.029839111698998347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,512,0.005208000126812193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,768,3072,0.0075662218862109715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,8192,0.18369244204627144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,128,0.004054222255945206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,8192,0.10901333226097955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,2048,512,0.022869333624839783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,8192,2.1379244062635636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,65536,0.4666124449835883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,128,0.009463999834325578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,8192,0.21797777546776664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,8192,1024,0.009135999613338048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,3584,0.010487111078368293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,2560,0.04020177655749851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,512,0.019711999429596793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,16384,2560,0.018223111828168232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,256,0.005419555637571547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,4096,0.018481777773963082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,8192,512,0.005518222020732031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,1024,0.0041306668685542215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,12288,0.26089954376220703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,512,10240,0.014340443743599786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,1536,0.011824000212881299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,1536,0.004570666700601578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,3584,0.030890666776233252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,12288,0.016390222642156813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,2560,0.02048711147573259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,128,4096,0.010805333654085795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,3072,0.030951112508773804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,256,0.026406221919589575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,5120,0.03280977739228143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,65536,0.23744000328911674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,3584,0.7226906882392036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,128,0.024978665841950312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,4096,20.94990709092882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,768,2560,0.00833422193924586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1024,5120,0.09134399890899658
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2560,1536,0.025224000215530396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,7168,0.03785422113206651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,2048,0.056427558263142906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,16384,0.06504444281260173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,5120,0.19780355029635957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,12288,0.2467555469936795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,3072,0.0750782224867079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,1536,0.04462577899297079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,128,0.002934222213096089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,7168,7168,0.024384000235133704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,5120,0.11320266458723281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,8192,0.03794666793611314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,5120,0.05894577503204346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,1024,0.24153510729471842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,8192,0.025081778566042583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2048,8192,0.046859555774264865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3584,768,0.06953244739108615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,7168,128,0.00573333352804184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,8192,1024,0.0653102199236552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,7168,7168,0.04400266541375054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,32,65536,0.06071022484037611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,10240,0.030710223648283217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,7168,768,0.006416888700591193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,512,0.05319111214743721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,2560,0.009463110731707679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,7168,12288,0.05172444383303324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,6144,0.16818133989969888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,65536,2.409664789835612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3584,6144,0.0336079994837443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,2560,0.007368889119890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,256,0.008619555168681675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,256,1024,0.056016001436445445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,10240,5120,0.03043022089534336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,12288,0.02804533309406704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,3584,0.0125591109196345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,12288,0.017512000269360013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,256,0.005133333305517833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,3584,0.018567111757066514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,2560,0.009147555463843876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,2048,0.007290666302045186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,3584,0.009576888548003303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,51200,2.3867314656575522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,3584,0.02741333345572154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,64,0.005231110999981563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,768,0.00564177789621883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,2560,65536,0.1820124387741089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,128,6144,0.0064311110311084324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,64,2560,0.005467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,6144,0.03018222252527873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,2560,256,0.0075048887067370946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,8192,8192,0.034134222401512995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,4096,768,0.006183111005359226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,5120,0.01919822229279412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,2560,0.035755554835001625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,256,0.023984889189402264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,1536,0.006772444479995304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,6144,65536,5.158420562744141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,1536,0.04754933383729723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,32,16384,0.0317191117339664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2048,10240,0.1227875550587972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,2048,0.006125333408514659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,1024,0.03129155437151591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,51200,0.01440711153878106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,64,0.0029102222373088202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,768,51200,0.03721066647105747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,512,768,0.01388711151149538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,1536,0.007719111111429002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,128,256,0.003273777870668305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,65536,0.07235466771655612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,2560,1536,0.007625777688291337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,7168,0.03905511233541701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,16384,512,0.06399289104673597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,4096,0.017338666650984023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,8192,1024,0.007628444168302748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,7168,0.06394222047593859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,12288,4096,0.020486222373114694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,256,0.006919999917348226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,512,768,0.013278221918476952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,6144,256,0.025182222326596577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,5120,0.01442399952146742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,1536,0.052645332283443876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,32,1536,0.00562755556570159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,16384,12288,1.468464003668891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,10240,16384,1.9813423156738281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,256,256,0.005429333282841577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,7168,0.48409779866536456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,6144,0.06261777877807617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,65536,0.14777510695987278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,768,0.009655999640623728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,2560,0.007178666690985362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,6144,0.006121777825885349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3072,10240,0.021320889393488567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,128,10240,0.013539555172125498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,128,0.003043555551105075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,2048,0.014116444521480136
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,65536,2560,1.1501803927951388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,512,0.025427555044492085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,3072,1.7442622714572484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,6144,5120,0.17268533176845977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1536,12288,0.15046132935418025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,51200,1.9350195990668402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,7168,0.09466489156087239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,256,0.003536000019974179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,3072,0.1436257759730021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,768,0.003705777641799715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,2048,0.005385777602593104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,5120,8192,0.019761777586407132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,65536,0.24580444229973686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,512,768,0.0038515557017591265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,3072,0.01536177761024899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1536,256,0.009071110851234859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,4096,0.022163555026054382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,256,0.008174222376611497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,2560,0.08439911074108547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,32,0.012304888831244575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,7168,3072,0.014489778214030795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,12288,10240,6.109817928738064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,4096,0.08406400018268162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,51200,2.0192008548312717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,7168,12288,0.036004443963368736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,10240,0.11324355337354873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,51200,5.719176822238499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,10240,0.09703822268380059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,768,64,0.006004444426960415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,8192,10240,0.39587288432651097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,128,0.002757333426011933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,2560,0.008522666990756989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,7168,0.011485333244005838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,1536,0.013099555340078143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,768,0.004586666822433472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,128,0.03717155588997735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,2560,0.035728000932269625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,65536,1.9204186333550348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,2560,0.058099554644690625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3584,3072,0.011674666570292579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,64,0.005380444642570283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2560,512,0.005938666562239329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,65536,17.972566392686634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,1536,0.01977599991692437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,256,0.0070568886068132185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,2560,1024,0.037371555964152016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,768,0.05611555443869697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,6144,0.12812888622283936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,768,51200,0.05796533160739475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,2048,0.027688889039887324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,51200,0.3261786566840278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,12288,0.899727079603407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,128,1024,0.006170666466156642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,4096,64,0.01460800071557363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,12288,12288,0.046889778640535146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,256,8192,0.05428444345792135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,4096,64,0.013677332964208392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,1024,0.094924443297916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,8192,0.17602399984995523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,16384,3584,0.142520891295539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3584,12288,0.047616001632478505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3072,2560,0.023870221442646448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,51200,3072,1.4724782307942708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,51200,0.07081955671310425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,7168,12288,0.03928622272279527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,256,0.010036444498433007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,6144,0.09328089157740276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,65536,0.6938551266988119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,128,1024,0.018597332967652213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1024,1024,0.012339555554919772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,512,0.02408711115519206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,51200,6144,1.495820469326443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2048,51200,0.10655022329754299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,4096,0.010781333678298526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,65536,256,0.011424889167149862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,6144,0.013097777962684631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,10240,0.01942133406798045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,3072,0.007646222081449296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,51200,0.16185600227779812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,512,3072,0.018048000004556444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,5120,0.013937777943081327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,12288,512,0.05962844689687093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,10240,2.6369696723090277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,768,5120,0.009621333744790819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,65536,3072,0.11011199818717109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,7168,0.9784399668375651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,5120,0.023559111687872145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,1536,10240,0.027083555857340496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,4096,0.010552888943089379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,10240,256,0.033506668276256986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,256,0.006811555474996567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,2048,0.13933332761128744
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,128,64,0.0029084444459941653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,256,2560,0.01651822196112739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,10240,0.02966133422321743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,6144,0.07254133621851604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,768,51200,0.43037954966227215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,8192,7168,0.0358551111486223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,1024,0.01107911103301578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,256,2560,0.020870221985710993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,256,1536,0.007710222568776872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,65536,4.808497958713107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,32,4096,0.015274667077594332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,12288,0.020222221811612446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,4096,3584,0.015792000624868605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,256,1536,0.006693333387374878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,16384,256,0.010503999888896942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,6144,3584,0.01760444376203749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,51200,0.5233102374606663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,1536,0.032623999648623996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,2560,0.008414222134484185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,128,1536,0.060870223575168185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,8192,1536,0.02221333318286472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,51200,4096,0.10894755522410075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,7168,0.011532444092962476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,7168,6144,0.9794443978203667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,5120,65536,0.20723644892374674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,65536,51200,1.7761529286702473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,4096,512,0.027162666122118633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,3072,0.04973777797487047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,3584,6144,0.013521778086821238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,256,1536,0.028427554501427546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,768,6144,0.03917955689960056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,6144,12288,0.04536177714665731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,4096,0.008476444416575963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,768,0.010984888507260216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,65536,0.1845724450217353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,3072,1024,0.00517511119445165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,512,65536,0.0702471137046814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,7168,0.24886933962504068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,64,0.013378666506873237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,32,0.007317333585686154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,8192,768,0.03691733214590285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,2048,16384,2.879886203342014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,65536,0.2351857821146647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,6144,0.006664888726340399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,5120,0.027148443791601393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,6144,0.45412264929877383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,12288,0.08560444249047174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,512,0.008941333327028487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,16384,6144,0.3492435614267985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,16384,0.02085688875781165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,8192,0.19260533650716147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,1536,0.007806222471925948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,768,0.04411022199524773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,8192,0.020828444096777175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,7168,6144,0.3034079869588216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,10240,768,0.06594400273429023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,512,0.014870221416155497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,8192,0.16656000084347194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,10240,3584,0.017861333158281114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,4096,32,0.01315111087428199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,51200,0.029264890485339697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,256,512,0.005486222190989389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,8192,0.3029164473215739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,65536,0.05616088708241781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,51200,8192,2.9672684139675565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,3072,0.021260443660948012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,32,51200,0.07418844434950086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,64,0.0038115556041399636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,8192,0.012826667063766055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,64,0.01401600076092614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,8192,768,0.010392888552612728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,256,0.005518222020732031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,51200,0.09217600027720134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,128,0.01184622198343277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,768,0.01591911084122128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,32,0.0258577780591117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,512,0.03053955568207635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,2560,0.008383110993438298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,64,3584,0.006784889019197888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,4096,0.016332444217469957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,5120,128,0.02416177756256527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,6144,0.10203022427029079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,256,0.17409066359202066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,10240,128,0.047107554144329496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,5120,3584,0.22028888596428764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,6144,0.010980444649855295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,128,128,0.007832888927724626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,4096,51200,0.06807466348012288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,5120,1536,0.009266667068004608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,2048,0.006638222270541721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,16384,6144,0.034487111700905695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,256,512,0.008839111361238692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,2048,0.026361778378486633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,2048,0.5609973271687826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,7168,2560,0.020941333638297185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,16384,7168,0.5544168684217665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,7168,0.022184888521830242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,32,10240,0.007556444240940942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,65536,2560,0.7481262418958875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,7168,64,0.02388800018363529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,768,4096,0.011810666984981961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,12288,0.019279999865425956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,2048,1024,0.00637511122557852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,2560,0.1374195549223158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,3072,0.015218666858143277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,6144,32,0.00291377781993813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,5120,2560,0.01257599973016315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,6144,0.15407289399041071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,256,0.06649955776002672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,3584,0.01612711118327247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,51200,51200,6.609823015001084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,16384,3072,0.20762844880421957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,256,0.008829333715968663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,2048,0.03561155663596259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,51200,0.7576515409681531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,3584,0.06726755698521932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,8192,0.04206044475237528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,5120,16384,0.03537155522240533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,6144,10240,0.5335689120822483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,1024,0.3350755638546414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,128,16384,0.024933333198229473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,3072,0.018920888503392536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,8192,0.01750666730933719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,10240,768,0.009184000392754873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,7168,8192,0.09371822410159641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,8192,0.01585155559910668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,1024,2048,0.024367999699380662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,51200,64,0.6827768749660916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,1024,0.006646222124497096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,768,0.062466667758093945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,10240,0.013424889081054263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,10240,1536,0.4473875363667806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,51200,1.3190524843004015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,51200,5.572076585557725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,10240,51200,0.27056090037027997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,64,768,0.0038542221817705366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,8192,7168,0.025674665967623394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,2560,0.061540444691975914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,4096,128,0.004382222063011593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,32,0.005186666631036335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,3072,4096,0.011377777490350934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,16384,32,0.039678222603268094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,5120,0.01201333353916804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,2560,0.008042666647169325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,10240,0.167048003938463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,32,0.017090666625234816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,3072,0.019542222221692402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,16384,0.9051919513278537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,8192,0.2946400112575955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,51200,64,0.004262222184075249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,65536,0.35051645172966855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,3584,0.16673244370354545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,10240,0.009553777674833933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,3072,0.014808000789748298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,12288,0.1822355588277181
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,128,5120,0.006558222075303395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,7168,1536,0.009523555636405945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,16384,0.04732622371779548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,4096,0.024795555406146582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,512,0.006069333189063602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,12288,0.07506577836142646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3584,64,0.013227555486891003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,16384,16384,0.12824800279405382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,65536,0.07153689199023776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,10240,0.04100799891683791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,65536,2048,1.5527351167466905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,768,0.013167999684810638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,12288,0.21955200036366782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,6144,0.04149688945876227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,4096,0.013255111045307584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,10240,0.027808000644048054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,3584,256,0.004744889007674323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,6144,0.027480888697836135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,4096,0.01372977760103014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,5120,6144,0.01628177695804172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,8192,0.013622221847375235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,51200,0.08098933431837294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,16384,4096,0.03702933258480496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,10240,0.026738666825824316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,1536,0.007996444072988298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,3584,128,0.006434666613737742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,4096,65536,0.08898488680521648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,10240,2048,0.019225777851210702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,768,0.010096888575288985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,2048,0.016001777516471017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,1024,7168,0.06191555658976237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,512,0.03511022196875678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,7168,64,0.10618666807810466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,5120,0.17794844839307997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,2560,0.009713778065310584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,128,0.03616799910863241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,65536,49.398332383897575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,128,0.004375111311674118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,2048,0.05043288734224108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2560,5120,0.015159999330838522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,128,0.005715555614895291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,51200,128,0.025920000341203477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,5120,256,0.0058764442801475525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,8192,3.533725314670139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,6144,0.01611822181277805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,12288,0.06553955872853597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,5120,0.07418844434950086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1536,5120,0.02409422232045068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,6144,0.06541955471038818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,2560,0.007711110843552484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,3584,0.03156177865134345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,1024,0.025229333175553217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,512,0.004977777600288391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,16384,0.019828443725903828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,1536,0.020163555939992268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,8192,0.30829066700405544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,12288,4096,1.4322399563259547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3072,12288,0.028558221128251817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,4096,0.02885955572128296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,51200,10240,0.24934845500522188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,5120,51200,0.5763875643412272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,16384,0.20691822634802925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,10240,0.026387555731667414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,12288,1.2189351187811959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,128,12288,0.020417778028382193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,16384,3584,0.40535643365648055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,7168,0.01588088936275906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,2560,0.036907556984159685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,64,0.0030826667530669104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,768,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,3072,0.034083555142084755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,3584,0.015585778488053216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,65536,0.1583040025499132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,6144,1024,0.02558311157756382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,6144,0.01960444450378418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,7168,0.2763502332899305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,1024,10240,0.01956355571746826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,8192,0.12536088625590006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,2048,5120,0.011119999819331698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,51200,64,0.17791467242770723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,10240,0.006334222025341458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,51200,51200,0.7127057711283366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,64,0.008074666890833113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,12288,0.018664000762833487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,7168,0.2034355534447564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,12288,0.3766382270389133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,12288,0.020802666743596394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,1024,0.0041422223051389056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,16384,16384,0.08787200185987686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2560,1024,0.014717333846622042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,12288,65536,0.23056001133388945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,3072,0.10325955682330662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2048,256,0.013535111314720578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,7168,64,0.03186577889654372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,5120,0.013839110732078552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,8192,0.03089511063363817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,512,0.00433955548538102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,12288,5120,0.02420799930890401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,4096,0.015842666228612263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3072,4096,0.07225155830383301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,7168,1536,0.009558222360081142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2048,16384,0.023349333140585158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,64,0.002699555622206794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,4096,0.011120888921949597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,2560,0.014825777875052558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,32,0.004208000169859992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,64,0.0027315556589100095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,64,0.0027066665805048416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,512,5120,0.23883912298414442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,64,0.005341333233647876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,64,12288,0.015603555573357476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,4096,512,0.018180444836616516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,1024,6144,0.05390222205056084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,12288,0.2337040106455485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,3584,0.1164826684527927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,32,0.002667555585503578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,768,768,0.015056000815497505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,51200,0.09375377496083577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,1024,0.006761777732107375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,2048,0.1414168940650092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,64,0.04059200154410468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,4096,0.011867555479208628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,512,256,0.005042666776312722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,12288,1536,0.01255733354224099
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,16384,0.05820444557401869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,768,768,0.005467555589146084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,6144,0.11210578017764622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,32,3072,0.008638222184446123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,2048,0.006360888895061281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,512,16384,0.1331226693259345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,512,0.01594044433699714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,512,0.00416711096962293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2560,5120,0.09118488762113784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,32,1024,0.008738666772842407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,32,0.002704888789190186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,3072,0.21786755985683867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2048,768,0.003642666671011183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,768,0.005958222266700532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,7168,0.021103999680942957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,768,0.0057795556883017225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1024,5120,0.032237334383858576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,51200,0.8681102328830295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,256,0.01237155579858356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,10240,0.1797057787577311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3584,6144,0.15830043951670328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,768,0.005952888892756567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3584,2560,0.051242666112052075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,512,8192,0.11999466684129502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,8192,0.022064000368118286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,4096,0.021439111895031396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,10240,0.0063573333124319715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,10240,1536,0.013357333011097379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,6144,0.130039996571011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,3072,0.017743999759356182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,1024,10240,0.015776889191733468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,6144,10240,0.18589156203799775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,2560,0.042945779032177396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,7168,0.1785128911336263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,32,12288,0.0345164438088735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,128,16384,0.03254400028122796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,65536,8.470807393391928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,7168,0.01425333321094513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,16384,8192,7.588079664442275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,128,3072,0.025669333007600572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,65536,32,0.08655377891328599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,7168,1536,0.008569777839713627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,7168,0.0063973334100511335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,256,0.007599111232492659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,7168,1024,0.03247377938694424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,1024,0.05257777704132927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,2560,0.01590844492117564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,128,0.00443111111720403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,8192,0.105968885951572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,16384,2048,0.018775999546051025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,512,1536,0.025229333175553217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,1024,0.04871911141607496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,4096,12288,0.3093786769443088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,512,256,0.005523555394675996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,5120,65536,0.19414044751061332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,1536,0.1541795598136054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,128,256,0.00793599999613232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,512,1536,0.024717332588301763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,512,0.010677333507272931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,768,0.00886399961180157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,65536,1.7440062628851996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,512,0.05252088771926033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,2048,256,0.018415111634466384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,2560,0.008549333446555668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,51200,6.315479278564453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2560,128,0.006817777951558431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,6144,768,0.007810666329330868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3072,2560,0.07854133182101779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,3584,2560,0.009669333696365356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3072,256,0.005697777701748743
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,4096,8192,0.053278221024407275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,32,0.020471110939979553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,65536,10240,0.7759528689914279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,768,8192,0.09803733560774062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,3584,512,0.005222222043408289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,6144,6144,0.22577243381076387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,3584,0.023570666710535686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,2560,0.01880088945229848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,16384,0.07458755705091688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,7168,512,0.004767111192146937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,16384,1.6169235441419814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,768,0.479656007554796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,64,1536,0.004950222041871813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,3072,0.007249777515729268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,64,768,0.0038746665749284956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,3584,0.0767235557238261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,51200,4096,11.249998304578993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,8192,0.07565689086914062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,1536,0.010832000109884473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,10240,65536,0.3681173324584961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,256,256,0.0042462220622433555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,4096,12288,0.10481422477298313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,4096,0.008844444321261512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,3072,0.0248364441924625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,12288,0.025881777207056682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,5120,0.024002666274706524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,12288,0.02386755579047733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,7168,0.034494221210479736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,2048,0.007965332931942409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,16384,0.11789066261715359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,65536,0.014922666880819531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,5120,0.2728533214992947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,5120,2048,0.013547555440002017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,2560,0.13212977515326604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,2560,0.006849777781301075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,16384,0.05954222546683418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,512,12288,0.05576444334454007
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,64,0.0028613333900769553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,2560,0.010470222267839642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,2048,64,0.006544000158707301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,32,32,0.00517866677708096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,12288,0.017438222964604694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,128,32,0.002609777781698439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,3584,0.01700622174474928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1024,3072,0.00813155538505978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,7168,1536,0.00924533357222875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,64,0.18881421618991426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,10240,0.04697688751750522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1024,256,0.006798222247097228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,6144,0.061936000982920326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,6144,2560,0.01109955542617374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,32,0.030699554416868422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,4096,0.0058871110280354815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,2048,3584,0.035678221119774714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,512,0.013263999587959714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3072,10240,0.03130044539769491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1024,16384,0.10675911108652751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,4096,1.3903769387139215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,128,0.007443555527263218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,3584,0.012773333324326409
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,128,0.0026986667265494666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,6144,0.0586951110098097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,16384,12288,0.37462488810221356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,64,0.00565155554148886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,1024,0.10038755337397258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,12288,2048,0.19235377841525605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,1024,0.006100444330109491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,1024,0.010439111126793755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,32,1024,0.010043555663691627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,4096,0.02760533326201969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1536,1536,0.010403555300500657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,1024,512,0.00544888898730278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1536,32,0.006636444479227066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,256,2048,0.007602666815121968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,512,32,0.0033662221911880705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,128,0.023150222169028387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,12288,2560,0.015680889288584392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,64,51200,0.2972639931572808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,16384,3584,0.0363377763165368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,768,8192,0.012476444244384766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,512,0.005572444448868434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,1024,0.00850133349498113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1536,7168,0.019357333580652874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,12288,1024,0.01014666673209932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,1024,0.00618577789929178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,1024,0.004837333328194088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,16384,2.133659574720595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,12288,7168,0.29971999592251247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2048,51200,0.17490577697753906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,256,0.02534755567709605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,768,51200,0.256111118528578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,12288,3584,0.024675556355052527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,2560,0.11876177787780762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,3072,0.009292444421185387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,16384,0.09442844655778672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,2560,0.03048000070783827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,4096,0.019116444720162284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,8192,6144,0.026183999247021143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,4096,0.009944889280531142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,512,0.009462222456932068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,512,2048,0.005543111099137201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,32,3072,0.021332444416152105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,4096,128,0.002944888960984018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,12288,0.0465555555290646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,256,0.01551555593808492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,32,256,0.007884444461928474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,8192,51200,1.3091386159261067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1536,5120,0.014313777287801107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,8192,0.03259999884499444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,16384,13.526931762695312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,32,0.0038417776425679526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,32,512,0.007490666376219855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,51200,0.7039430936177572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,16384,0.25565067927042645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,512,0.003527111063400904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,65536,0.09575021929211086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,768,0.0063573333124319715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,3072,0.006715555571847492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,7168,0.3878204557630751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,512,0.008957333034939235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,6144,0.017322666115230985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,2048,0.030605332718955144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,5120,0.06748355759514703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,51200,10.971358405219185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,128,256,0.005212444398138259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,512,0.035368889570236206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,256,0.009298666483826106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,7168,8192,0.03357333276006911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,51200,0.3567795488569472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,6144,1.3901928795708551
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,512,0.005713777823580637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,8192,0.01824977828396691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,10240,16384,0.06007644202974108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,4096,0.010026666853162978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,3584,0.008266666697131263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,32,0.0031235555393828284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,12288,0.05905510981877645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,3072,7168,0.016919111212094624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,1024,10240,0.019792889555295307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,4096,0.011911999848153857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,768,0.003689777933888965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,256,3072,0.00850044439236323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,1024,768,0.016136000553766888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,5120,0.7106213569641113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3072,4096,0.013171555267439948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,1536,0.033173332611719765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,65536,10.551710340711805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,10240,0.14654755592346191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,4096,0.022106667359670002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,8192,0.02199111051029629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,7168,2560,0.013170666992664337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,4096,0.06403822369045682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,64,0.006616000086069107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,16384,0.39427910910712344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,1024,0.007340444458855524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,51200,32,0.0042133331298828125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,51200,5.596381293402778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,7168,16384,0.0684151119656033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2048,2560,0.042540444268120654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,768,0.0063164445261160536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,2560,3072,0.010010666317409938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,256,0.0036026665733920205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,3072,0.10197155343161689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,65536,0.032687998480266996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,65536,3072,1.7135413487752278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,7168,0.020131554868486192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,51200,0.032484445306989886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,10240,10240,0.035571555296579994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,512,0.004949333353175057
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,12288,3584,0.023950222465727065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,3584,0.019471999671724107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,8192,16384,0.765135129292806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,10240,128,0.008092444803979661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2560,3072,0.009182222187519073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,3584,0.011617778076065911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2560,6144,0.10760000016954209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,10240,64,0.024884444144037034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,51200,16384,0.4428577688005235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,8192,16384,0.04683377676539951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,7168,512,0.042711112234327525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,51200,2560,0.19071555137634277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,7168,0.16898577743106416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3072,3072,0.012784889174832238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,32,0.06302399767769708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1536,6144,0.01367555558681488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,12288,0.21384621991051567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2560,65536,0.6198568873935276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,128,0.00554577757914861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1024,3584,0.033210668298933245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,3072,0.008325333396593729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,2560,16384,0.23481776979234484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3584,10240,0.02275288932853275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,512,51200,0.1524426672193739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,7168,65536,0.14322400093078613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,768,0.00962222201956643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,6144,256,0.004808889081080754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1536,512,0.005984000033802456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,32,0.0026231110095977783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,7168,512,0.008981333010726506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,64,0.0028400001012616684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,65536,0.4587048954433865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,128,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1536,5120,0.010089777410030365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,256,2048,0.028841776980294123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,5120,2048,0.009233777721722921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,10240,12288,0.39447021484375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,64,10240,0.07579021983676486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,65536,0.014704888065656027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,128,1024,0.011908444265524546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,7168,3584,0.05923644701639811
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,51200,1.3665528827243383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,64,3072,0.02639377779430813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,256,0.003927999900446998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,5120,0.04646311203638712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,64,3584,0.007491555478837755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,51200,1.3540835910373266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,6144,0.014444443914625378
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,65536,10240,8.299936082628038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,8192,0.05163733495606316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,3584,10240,0.051121777958340116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,6144,0.012100444071822695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,512,128,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,8192,0.03349155518743727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,8192,7168,0.04050755500793457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,3584,0.007856888903511895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,1536,0.014968888627158271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,64,5120,0.01239733315176434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,65536,0.06180266539255778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,512,8192,0.011554666691356234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,512,12288,0.19301422437032065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,4096,0.03167111012670729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,32,3072,0.008264888491895463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3584,2560,0.012598222328556908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,3072,0.026501332720120747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3584,1024,0.004920888692140579
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,10240,0.08967466486824884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,3584,16384,0.04014399978849623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,16384,0.009535111486911774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,1536,0.05193688803248935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,128,0.004849777867396672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,64,10240,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,768,10240,0.016001777516471017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,768,1024,0.008787555827034844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,4096,12288,5.076294369167752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,128,0.003522666792074839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,51200,0.29152711232503253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3072,1024,0.006256888724035687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,5120,51200,1.5683555603027344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,16384,8192,0.05351466602749295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,65536,1.0134560267130535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,768,1024,0.012311110893885294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,256,4096,0.04833244283994039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,1024,0.028911110427644517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,1024,0.00981244444847107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,64,8192,0.008984888593355814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,3584,0.07320266962051392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,12288,2048,0.3010302119784885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,8192,5120,0.02089244458410475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,7168,8192,0.4635964499579535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,2048,0.058394663863711886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,16384,5.282937791612413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,4096,3072,0.019766221443812054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,128,6144,0.01372977760103014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,6144,51200,0.15065156088935003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,2048,0.07814488808314006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,5120,65536,3.5935020446777344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,4096,0.011549333731333414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,4096,3072,0.0258586671617296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,6144,32,0.008292444050312042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,65536,256,0.023183110687467787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,128,10240,0.08761866887410481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1024,4096,0.009929777847396003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,2560,0.023896889554129705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,64,10240,0.037363555696275495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,4096,0.008643555144468943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,4096,2560,0.02183733383814494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,2560,128,0.011600000162919363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,16384,4096,0.02208799951606327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,8192,8192,0.7355804443359375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,8192,0.008202667037645975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,128,0.012364444633324942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,12288,12288,0.8890470928615994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,6144,1536,0.010864888628323873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,256,0.27010220951504177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,1536,0.13682222366333008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,16384,1024,0.03122488988770379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,16384,0.0894391139348348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,5120,0.021928000781271193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,12288,0.266881783803304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,4096,0.019516444868511625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,64,0.006117333140638139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3072,64,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3072,16384,0.03109955456521776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,10240,0.017144888639450073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,512,0.03390311201413473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,8192,51200,0.14987999863094753
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,65536,1024,0.039353777964909874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1024,128,0.004139555411206352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,4096,0.010647111468844943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,4096,0.008976000050703684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,65536,512,0.025379555092917547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,7168,512,0.030125333203209773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,256,51200,0.0787866645389133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,1024,0.004966222163703707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,6144,0.03978044456905789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,12288,1024,0.0807831088701884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,12288,0.05115555392371284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,32,0.00665244460105896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,32,0.007511999871995714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,10240,0.055367999606662326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,512,512,0.007400888535711501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,64,51200,0.049192888869179614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,512,0.015419556034935845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,12288,0.18477866384718153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,768,0.0036311112344264984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,16384,768,0.009451555709044138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,12288,0.02419822249147627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,51200,64,0.024255999260478552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,1536,0.0054560001525614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,12288,0.06266578038533528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,256,0.008231110870838165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,768,0.029292444388071697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,4096,0.27315110630459255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,65536,0.2476310994890001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,3584,0.12367110782199436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3072,128,0.004469333423508538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,65536,0.14830933676825628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,10240,0.04650933212704129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,6144,0.2194275591108534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,10240,6144,0.02421244482199351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,12288,2560,0.021180444293551978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,4096,65536,0.15444178051418728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2048,512,0.005941333456171884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,1536,0.011996444728639392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,4096,0.016719111137919955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,1024,0.017261332935757108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,128,4096,0.007974222302436829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,128,0.003003555453485913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,2048,0.02204088866710663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,4096,0.08025066720114814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,7168,0.056997332308027476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,8192,4096,0.2214204470316569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,1536,3072,0.007840000092983246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,16384,0.10372977786593968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,7168,0.026718222432666357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,1536,0.05236088898446825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,10240,1.1140658060709636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,8192,12288,0.4701804584927029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,51200,1024,0.21548177136315239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,4096,0.00925866680012809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,512,0.03994311226738824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,7168,0.020092444287406076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,2048,3584,0.00997866690158844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2048,7168,0.013645333548386892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,1024,0.033293333318498396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,7168,0.015464888678656684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,7168,12288,0.09812000062730576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,10240,0.049570666419135205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,16384,256,0.03574311070972019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,16384,10240,0.1047368910577562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,65536,3584,0.07008355855941772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,1024,0.0058808889653947614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,256,3584,0.03377422359254625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,7168,0.046875556310017906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,2560,3584,0.13903289371066624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,8192,64,0.027663111686706543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,6144,65536,0.17460178004370797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,5120,4096,0.011786667009194693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,256,3072,0.007497777541478475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,16384,0.057757331265343555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,65536,0.12149777677324082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,256,768,0.0036115555299652945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,2048,0.006543111056089401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,5120,0.06282133526272245
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,51200,16384,0.6988222334120008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,7168,0.09946044286092122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,5120,0.022113778524928625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,8192,32,0.026703110999531213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,768,3072,0.008378666308191087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,12288,10240,0.06510844495561388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,1024,1536,0.0070497774415545994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,32,2560,0.008671999805503422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,16384,0.03702488872740004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,8192,12288,0.08402133650249904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,10240,65536,3.3195866478814016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1024,7168,0.01780533293883006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,128,0.0031893334041039148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,768,3584,0.054494222005208336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,6144,0.00775733341773351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,128,1024,0.003961777935425441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,5120,0.007369777394665613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,12288,0.5666071044074165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,768,0.005379555539952383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,51200,1536,0.02923822071817186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,65536,0.1069342229101393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,65536,7168,18.041673448350696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,3584,0.01850133306450314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,8192,512,0.008395555946562026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,768,0.006677333265542984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,8192,0.028566221396128338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,51200,2560,0.6743582089742025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,8192,0.011047110789351992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,128,0.003167111012670729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,128,0.05147822366820442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,8192,3584,0.0678408874405755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,12288,51200,0.8033715354071723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,256,0.03216088811556498
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3072,512,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,64,0.003141333245568805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,10240,2048,0.013628444737858243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,51200,8192,1.954204347398546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2048,65536,0.07687466674380831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,4096,0.07678133249282837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,10240,512,0.005966222120655908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1536,3072,0.035374220874574445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,4096,0.07288533449172974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,2048,0.07330755392710368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,5120,7168,0.27757244639926487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,3584,1536,0.00740444411834081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,65536,8192,4.883335961235894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,6144,0.02525866693920559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,3584,1536,0.03204710947142707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1024,512,0.006764444626039929
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,51200,8192,1.9305118984646266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,51200,2048,0.6654026773240831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,7168,2.261399163140191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,6144,3072,0.016121778223249648
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,768,3072,0.017507554756270517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,5120,0.07088622119691637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,512,2560,0.009076444639099969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,6144,16384,0.5839804543389214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,8192,128,0.0337413317627377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,32,0.003993777765168084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,512,8192,0.015198222464985318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,7168,16384,0.5885698000590006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,65536,256,0.015558222929636637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,1536,0.006863111009200414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,2048,0.07773511277304755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1024,2560,0.0070088886552386815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,2560,256,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,256,0.07453244262271456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,8192,16384,0.09920622242821588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,1024,0.004701333327425851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,128,2560,0.01834488908449809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,32,0.1616764465967814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,32,768,0.0036346668170558084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,65536,2560,0.32582222090827095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,7168,0.18795643912421334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,256,0.004707555390066571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,4096,16384,0.027220444546805486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,65536,0.4764791064792209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,1536,0.011612444288200803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,32,3072,0.006687110910813014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2560,64,0.013776889277829064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,512,0.009288888838556077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,128,2048,0.007747555772463481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,12288,128,0.009363555245929295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,4096,0.024829333027203877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,768,0.0069066666894488865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3584,6144,0.01423377792040507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,12288,0.5584195454915365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,12288,0.10339022345013088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,512,0.00793511089351442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,8192,0.040165334939956665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,3584,0.06473777691523235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,4096,0.036758220858044095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,7168,0.03638844357596503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,5120,10240,0.5554506513807509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,1024,51200,0.07148177756203546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,2560,256,0.00721066693464915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3584,1536,0.4974755711025662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,10240,32,0.02956355611483256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,2560,0.017887999614079792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,768,0.018241778016090393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,5120,256,0.005897777775923411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,12288,0.060770663950178355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,65536,2560,0.05584000216590034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,5120,0.03486488925086127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,256,2560,0.006697777658700943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,3584,0.017124444246292114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,8192,0.02453600035773383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,2048,0.01610311037964291
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,7168,0.03442666596836514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,7168,2.4166764153374567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,3584,12288,0.02474044428931342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,10240,0.08130400048361884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,768,0.021027555068333943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,5120,2048,0.011132444772455426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,128,0.0429422226217058
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,2048,0.01219288922018475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,4096,4096,0.013995556367768182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,8192,3072,0.015603555573357476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,8192,1024,0.007279111279381647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,512,256,0.004353777815898259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,65536,0.14492621686723497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,768,2048,0.02607911162906223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,8192,0.04420977830886841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,12288,2560,0.04705955584843954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,768,768,0.009129777550697327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,16384,5120,0.02997866604063246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,1536,0.011848889291286469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,51200,16384,0.40120177798801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,65536,0.06907644536760119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,8192,7168,0.15458222230275473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,6144,7168,0.09391199880176121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,256,51200,0.3139102194044325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,12288,0.04398488998413086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,64,7168,0.012475555141766867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,5120,0.04219111137919956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,51200,0.17695021629333496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,65536,0.7319591310289172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,7168,4096,0.20795288350847033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,32,16384,0.008397333323955536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,8192,0.3738826645745172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,12288,0.4448497560289171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,3072,0.04804355568355984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,128,0.009024000002278222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,5120,0.47397957907782656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3072,3584,0.012264000044928657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,3072,6144,0.13694577746921116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,5120,0.17430222034454346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,16384,0.05797421932220459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,6144,0.009715555442704095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,65536,1024,0.025916443930731878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,256,8192,0.013578666581047906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,5120,0.010008888940016428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,128,0.005430222385459476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,5120,128,0.0032026666320032547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,51200,32,0.09928533103730942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,7168,0.029990222718980577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,512,0.015302222636010913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1024,768,0.005777777896987067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,7168,0.25911466280619305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,7168,0.031403554810418025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,4096,6144,0.013910222384664746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,51200,0.3140790992312961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,128,0.0030542222989930045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,6144,0.00996088898844189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,768,0.09796088933944702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,65536,7168,0.12662933932410345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,768,0.048150221506754555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,10240,8192,0.12251733409033881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,10240,12288,0.04294133186340332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,16384,5120,0.04278488953908285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,6144,256,0.004824000100294749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,128,1024,0.00425511101881663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,4096,0.010982222027248807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,4096,0.5918266508314345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,2560,0.016152888536453247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3584,51200,1.5936995612250433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,65536,1024,0.19294222195943198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,1536,16384,0.028970665401882593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,4096,16384,0.028628445333904688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3584,10240,0.026571555270089045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,4096,0.11093066798316108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,3584,0.01778133379088508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,1024,7168,0.010271111296282874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,256,0.21253600385453966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,51200,5120,0.7771209081013998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,6144,10240,0.023648000425762598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,512,7168,0.03609866566128201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,128,512,0.004464888738261329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,64,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,1536,8192,0.02199733257293701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,768,0.02797422144148085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,64,16384,0.11674666404724121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,3584,0.01200888885392083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,5120,0.587208006117079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,5120,0.09114399883482192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,12288,7168,0.03257333238919576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,16384,0.022522666388087805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,5120,0.008192888564533657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,64,0.007184888753626082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,2560,0.006060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,6144,0.017129777206314933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,64,0.19386755095587838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,1536,0.007319110963079665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,51200,0.1948000060187446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,768,3072,0.007583111524581909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1536,2048,0.005561777700980504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,10240,0.025203555822372437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,256,0.011617778076065911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,3584,0.007816000117195977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,2560,0.08080622222688463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,4096,0.019907555646366544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,51200,2048,0.06147999895943535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,32,2048,0.007055111229419708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,51200,2560,0.08173955811394586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,12288,65536,5.676286909315322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,2560,0.0683857798576355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1536,7168,0.05204000075658163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,4096,0.04054222173160977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,768,0.05367288986841837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,512,0.0036480000449551474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,32,0.005260444349712796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,1024,0.019140443868107267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,512,0.005414222263627582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,10240,0.2793040010664198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,2560,0.011218667030334473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,5120,0.3055608802371555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,8192,128,0.03244977858331468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,2560,0.13004444705115423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,6144,5120,0.028106666273540918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1024,4096,0.008236444658703273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,65536,8192,0.16215377383761936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,512,0.00849422232972251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2048,512,0.05730488565233019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,8192,16384,0.6631644566853842
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,7168,0.08376977841059367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,4096,3584,0.013718222578366598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,6144,4096,0.014101333088344999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,256,0.005426666802830166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,2560,32,0.029622223642137315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,12288,0.15336622132195368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,8192,6144,0.4699866506788466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,3584,1536,0.009655111365848118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,6144,0.024149333437283833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,32,0.2311262289683024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,16384,7168,0.04428355561362373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3584,128,0.05637866920895047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,51200,0.05586222145292494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,7168,2560,0.012062222593360476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,5120,3072,0.01352355546421475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,7168,0.028132445282406274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,2560,0.4536693361070421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,3072,0.0599271125263638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,256,768,0.015505777464972602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,16384,0.04997511042488945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,512,256,0.006707555717892117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,10240,0.06096711423661974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,768,256,0.004965333475006951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,256,768,0.023450666003757056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,768,0.02963288956218296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,6144,0.006026666611433029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,65536,0.16295111179351807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,8192,0.17698933018578422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,512,0.03378666771782769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,65536,0.07238311237759061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,512,0.006863111009200414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,6144,0.3331475522783068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,65536,0.10164444314108954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,5120,768,0.006118222243256039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,32,0.045231110519833036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1536,51200,0.8627120123969184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,7168,1536,0.009713778065310584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,1536,0.03776622149679396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,4096,0.011247110863526663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,256,0.010668444136778513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,8192,6144,0.3841502136654324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,1536,5120,0.05405066741837395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,128,65536,0.015863999724388123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,3584,7168,0.01926488843229082
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,12288,2560,0.015325332681337992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,51200,8192,0.1163511143790351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,5120,0.13489333788553873
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,8192,4096,0.25656265682644314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,6144,768,0.006463999963468975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,12288,6144,0.33449776967366535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,8192,5120,0.019646222392718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,16384,6144,0.04784977767202589
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,16384,3.298678292168511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,65536,4096,0.1340577734841241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2048,10240,0.020584000481499564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,3584,8192,0.021185777253574792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,2048,0.00813688917292489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,1024,12288,0.05617333120769925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,768,7168,0.011154666543006897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,256,6144,0.01000355515215132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,1536,0.006411555740568373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,64,7168,0.006517333288987477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,16384,3.8988265991210938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,2560,768,0.018811555372344125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,8192,0.012607110871209038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,12288,0.039080000585979886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1536,4096,0.018215111560291715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,512,64,0.0034168888297345904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,2560,0.04605866803063286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,6144,5120,0.06719466712739733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,32,0.006600889066855113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,16384,0.1632844474580553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,51200,0.05670577949947781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,768,0.026458667384253606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,128,5120,0.038350221183564924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,65536,128,0.08210933208465576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,128,0.05033955640263028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,512,0.007125332951545715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,256,0.010063110954231685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,7168,0.24211645126342773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,5120,0.09266755315992568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1024,32,0.0029315555261241067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,6144,0.22757244110107422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,768,0.0047973332305749255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3584,512,0.017143110434214275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,65536,0.24015466372172037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,512,0.009397333694828881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,128,0.0029146667155954572
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,51200,8192,0.12658045026991102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1024,32,0.0054551110499435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,51200,0.06842666864395142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,51200,128,0.09507111046049331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,10240,0.04958044489224752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,4096,0.019751111666361492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,1536,1024,0.005867555737495422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,2048,0.013305777476893531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,16384,3072,0.13880088594224718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,7168,7168,0.02332177758216858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,3584,0.007825777762466008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3072,512,0.0068897778789202375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,10240,0.055027554432551064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,128,0.018882667024930317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,256,256,0.003277333246337043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,4096,0.009707555174827576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,256,0.030057777961095173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,2048,0.027495111028353374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,256,3072,0.007448888487286038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,65536,512,0.018329777651362948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,512,65536,0.06632888979381986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,7168,0.13210844993591309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,65536,1.7242373360527887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,2048,0.006775110960006714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,6144,0.1708408859041002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,10240,7168,0.03144444359673394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,8192,1536,0.08994933631685044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,8192,0.19007555643717447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,8192,7168,0.3915244473351373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,12288,0.011556444068749746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,16384,4096,0.15288710594177246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,768,0.023873777853118047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,512,128,0.00676711110605134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,1536,12288,0.01814399990770552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3584,12288,0.020593777298927307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,768,0.0058293334311909145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,65536,0.08875733613967896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,65536,7168,4.298951043023004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,51200,7168,1.368686252170139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,1536,0.0068986668354935115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,128,16384,0.018939556346999276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,32,0.009891555541091496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,1536,1024,0.004027555386225383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,10240,1536,0.015242666006088257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,768,2048,0.015418666932317944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,4096,0.01569688816865285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,51200,3584,1.4112603929307725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,10240,0.03578311204910278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3584,16384,2.8824168311225047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,8192,0.29384178585476345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,256,768,0.0166657782263226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,16384,0.10219999816682603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,1536,0.04523555437723795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,7168,0.024731554918819006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,16384,12288,0.09549244244893391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,8192,0.02270488937695821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,3072,8192,0.14246843920813665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,3584,5120,0.09128089083565606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,7168,256,0.01777866648303138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,5120,0.02421066661675771
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,1024,0.0064328888224230866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,128,51200,0.014143110977278816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,3584,32,0.00282222218811512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,256,0.0029688889367712867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,5120,0.015754666593339708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,51200,8.179432339138454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,5120,0.14200088712904188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,256,5120,0.038520889149771795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,8192,0.012653333445390066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,256,768,0.004486222234037188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,65536,16384,0.28513601091172963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,512,2560,0.011491555306646558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,3584,51200,0.08303644259770711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,2048,32,0.005453333258628845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,256,0.0057884446448749965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,12288,10240,0.07468977901670668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,65536,0.1260346704059177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1536,51200,0.6191146638658311
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,6144,0.07078933053546481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,7168,0.006903999795516332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,4096,0.007855999800893996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,32,0.004566222015354368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,1024,0.008866666919655269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,512,0.006704888823959563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,4096,64,0.005221333354711533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,3072,0.006518222391605377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,2560,0.20439467165205213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,3584,768,0.03750400079621209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,1536,0.00721866637468338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,65536,0.10491377777523464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,65536,3584,0.16878133349948457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,12288,512,0.005853333406978183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,65536,2.7190284729003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,16384,5120,0.041835556427637734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,51200,1.9983253479003906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,5120,0.008447999755541483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,4096,512,0.0040737779604064096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1024,256,0.005274666680230035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,768,0.005838222387764189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,4096,65536,1.2253973219129775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,10240,0.023108444280094568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,10240,256,0.028780443800820246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,12288,0.366575108634101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,3584,0.013638222383128272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,768,12288,0.11896355946858723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,8192,0.03692355420854356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,512,3584,0.01534933348496755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,768,128,0.006768888897365994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,2560,5120,0.011767110890812345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,256,0.005534222142563925
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,4096,1536,0.008413333031866286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2048,7168,0.03147111005253262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,7168,256,0.003419555516706573
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,8192,10240,0.03469511204295688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,5120,0.10801866981718276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,768,2560,0.016436444388495553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,8192,0.029653333955340918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,768,0.033478223615222506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,65536,2.625932481553819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,256,256,0.003216000066863166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,1024,0.006893333461549547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,768,0.024127999941507976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,12288,8192,0.07095288568072848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,5120,0.013567999833159976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,8192,6144,0.38513776991102433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,16384,5120,0.029984000656339858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3584,1024,0.02830577890078227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,3072,0.01257777793539895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,5120,0.013090666797426013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,3072,6144,0.01500266624821557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,5120,2560,0.010245333115259806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,256,2560,0.0074337778819931885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,512,0.07673688729604085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,65536,1024,0.025015999873479206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,1024,0.020364445116784837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,4096,256,0.02169688873820835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,768,256,0.005079999979999331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,5120,0.25367821587456596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,768,0.11693510744306777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,2560,65536,0.25336355633205837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,3072,0.01739199956258138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,4096,0.01236977759334776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,51200,0.716440889570448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,512,256,0.004794666750563515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,8192,0.014528888795110913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,512,51200,0.7847537994384766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3584,65536,0.6549955474005805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,16384,12288,0.832450654771593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2048,1536,0.028718223174413044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,32,0.0036844445599450003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,5120,0.02271733350223965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,256,7168,0.03783199853367276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,32,0.007332444190979004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,51200,12288,1.961316426595052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,51200,1024,0.03127911024623447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,3072,0.010791111323568555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,65536,0.1456435521443685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,51200,16384,0.42471999592251247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,1536,6144,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,4096,512,0.005959110955397288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,2048,128,0.0047280001971456735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,12288,64,0.06292711363898383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,5120,10240,0.23268532752990723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,128,51200,0.053098665343390576
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,10240,128,0.02810933192571004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,768,1024,0.0053288886944452924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,1024,256,0.005592888842026393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2560,2048,0.008626666333940294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,16384,32,0.031390223238203264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,5120,0.24024621645609537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,5120,4096,0.0332293344868554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,768,12288,0.018735999862353008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,16384,65536,11.900304158528646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,512,0.007471111085679796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,1024,0.007459555235173967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3072,64,0.004352000024583605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,16384,0.3011564413706462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,51200,3072,0.10679822497897679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,8192,6144,0.04090222385194566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,3072,0.09255288706885444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,1024,3072,0.009141333401203156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,64,1536,0.009867555565304225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,6144,512,0.005550222264395819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,4096,0.07896444532606336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,1536,0.03470755616823832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,256,0.0030177777840031516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,12288,7.366340637207031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,7168,51200,2.79418585035536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,8192,2048,0.01669066647688548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,2560,12288,0.12598578135172525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,16384,6.158608754475911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,6144,51200,0.1514328850640191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,4096,0.09702399704191421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,4096,0.09774488872951931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,64,128,0.005072000126043956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,1536,0.17229866981506348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,4096,0.019242667489581637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,51200,512,0.051691555314593844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,2048,0.007340444458855524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,12288,0.11856355932023789
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,3584,0.2671813435024685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,5120,0.013960889644092984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,1024,0.01088177743885252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,1024,0.007884444461928474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,7168,3.1246585845947266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,4096,7168,0.18769689400990805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,3584,0.04302933481004503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,3584,0.1076328886879815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,51200,32,0.2683724562327067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,1536,0.007082666787836287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,256,0.005756444401211209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,3072,0.014008000493049622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,1536,51200,0.04734488990571764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,10240,4.3732961018880205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3072,2048,0.05507111218240526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,256,4096,0.00959022260374493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,12288,10240,0.0789688891834683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,65536,0.07679910792244805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1536,6144,0.10115377770529853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,7168,16384,0.05001688996950785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,2560,0.014648000399271647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,768,2048,0.00867733359336853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,6144,6144,0.4089511235555013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1024,1536,0.03282666537496779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,10240,0.33333955870734316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,64,16384,0.007449777589903937
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,1536,0.017479111750920612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,32,128,0.002921777880854077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,7168,64,0.005884444548024072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,1024,3584,0.04347999890645345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,1536,16384,0.018016889691352844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,64,64,0.002564444517095884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,1536,3072,0.008027555214034187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,16384,0.1891608900494046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,3584,0.38279112180074054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,5120,0.014961777461899651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,512,0.003607999947335985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,512,0.025864000121752422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,5120,6144,0.2541573312547472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,2560,2560,0.007976888782448238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,128,256,0.007792889244026608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,512,32,0.002809777855873108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,12288,1024,0.010201777848932479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,3584,1024,0.04562933246294657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,65536,0.4221004380120172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,3072,0.00959911114639706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,2560,5120,0.015588444140222339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1024,6144,0.014567999376191033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,4096,256,0.0057706667317284485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,5120,0.040229333771599665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,10240,2560,0.015878222054905362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,51200,0.1307848824395074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,768,7168,0.030982222821977403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,128,10240,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,12288,0.02020355562369029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,4096,0.006730666591061487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,12288,0.02909066610866123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,3072,0.02067733307679494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1024,512,0.003428444473279847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,768,65536,0.5279351340399848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,2048,0.008392000363932716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,1024,0.008412444757090675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,128,0.0041599998043643106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,12288,2560,0.014887111054526435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,1024,0.004453333301676644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,12288,0.1134035587310791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,2560,0.007453333172533248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,512,16384,0.009521777431170145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,65536,128,0.12577422459920248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,10240,4096,0.17709155877431235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,4096,0.13467021783192953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,16384,0.23711821768018934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,16384,0.05242222216394213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,8192,128,0.024365334047211543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,16384,0.0498924454053243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,256,5120,0.023726221587922838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,64,0.002569777684079276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,8192,4096,0.02036711076895396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,512,0.004335999902751711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,128,256,0.0028924445311228433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,6144,10240,0.3541999922858344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,256,0.0069208890199661255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,1536,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,6144,3072,0.12404089503818089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,16384,1536,0.015104000767072042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,64,64,0.0028906667398081887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,1536,768,0.010491555763615502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,3072,12288,0.019838222199016146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,3072,8192,0.03784088955985175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,768,4096,0.006773333168692059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,3584,0.18081155088212755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,51200,19.708507961697048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,65536,8192,0.30405155817667645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,5120,256,0.003306666596068276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,128,256,0.00545866663257281
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,32,512,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,2048,64,0.004265777766704559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,2048,0.0236435549126731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,12288,0.01906844476858775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,12288,0.09387377897898357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,512,768,0.006278222219811545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,65536,128,0.10647910833358765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,16384,7.669464959038629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,3584,0.030995554394192163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,51200,2048,0.18632888793945312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,128,0.005199111170238919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3584,768,0.029350221157073975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1536,65536,7.5877532958984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,12288,4096,0.49435112211439347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,1536,0.010240000155236986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,3584,0.015595555305480957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,512,0.015856888559129503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3584,3584,0.03477866782082452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,8192,0.022356443934970435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,51200,3072,0.10185155603620742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,64,65536,0.07607644134097628
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,12288,0.03340177734692892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,51200,2.4663590325249567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1536,8192,0.011725333001878528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3072,1024,0.022277333670192297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,32,0.002581333327624533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,256,10240,0.03863199883037143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3072,12288,0.019518221418062847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,6144,64,0.0053582220441765254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,1024,1536,0.007060444189442529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,6144,2560,0.4757946862114801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,51200,1.3297369215223525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,16384,0.28905510902404785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,1536,0.02477688921822442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,8192,5120,0.02302933401531643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,51200,4096,0.06893155309889051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,5120,0.022242666946517095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,128,0.0074177781740824384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,768,16384,0.034041778908835515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,512,4096,0.03167288833194309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,3072,65536,1.5466515223185222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,7168,0.20449154906802705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,12288,0.018719111879666645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,5120,128,0.006986666884687211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,256,64,0.003540444291300244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,4096,12288,0.03476000163290236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,6144,4096,0.016985777351591323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,12288,1536,0.011507555842399597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,4096,0.011003555523024665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2048,4096,0.01148977792925305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,32,0.010580444501505958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,12288,1024,0.026719111535284255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,7168,3072,0.14254221651289198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,512,0.005410666680998272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3072,3584,0.010501333408885531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2048,64,0.009325332939624786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,65536,1.150059594048394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,768,0.03823466764556037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,12288,0.12577333715226915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,65536,0.127384000354343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,1024,0.004791111167934206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,128,65536,0.36640087763468426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,12288,0.02552266584502326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,16384,2048,0.020953777763578627
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3584,3584,0.14944533507029215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,16384,768,0.012825777961148156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,65536,5120,0.16495288742913142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,2560,4096,0.12620622581905788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,4096,6144,0.017816889617178176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,128,0.0027075554761621687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,5120,7168,0.026872888207435608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,768,3584,0.04915644394026863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,512,0.007239110767841339
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,7168,5120,0.11240000194973415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,2048,0.00776800016562144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,2560,2048,0.05849244197209676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,7168,128,0.004331555631425646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,1024,0.02276177704334259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,512,64,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,4096,0.042113777663972646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,7168,0.07348799705505371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,3584,65536,0.14866933557722303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,7168,0.5629875395033095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,3584,0.010870222416188983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1536,512,0.003858666867017746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,51200,2560,0.15183999803331164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,2560,0.24460800488789877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,16384,0.023620444867346022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,1536,0.4021342330508762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,16384,0.02437777817249298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,5120,512,0.0052844443255000645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,1024,128,0.00812177773978975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,768,0.006577777779764599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,256,256,0.005072000126043956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,7168,0.3922746711307102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1024,7168,0.028739555014504328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1536,1536,0.00830666638082928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,7168,0.14239822493659124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,1536,0.05680888891220093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,128,0.21558133761088052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,8192,3072,0.015245333313941956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,768,0.0057555557125144536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,7168,256,0.003228444399105178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1536,6144,0.07581866449779935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,1536,0.017228444417317707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,7168,0.012919999659061432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,6144,3584,0.012924444344308643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,12288,16384,0.08019377787907918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,16384,2560,0.018921777606010437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,4096,0.08732710944281684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,12288,128,0.031231999397277832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,768,0.11826222472720677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,12288,0.015554666519165039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,32,0.002593777866827117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,65536,768,0.14836266305711535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,51200,128,0.018029333816634286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,5120,2048,0.05636088715659248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3072,1024,0.02327999969323476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,51200,1.451511065165202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,7168,0.009393778112199571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2048,6144,0.05307022068235609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,32,1536,0.005029333134492238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,3584,8192,0.02676711148685879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,65536,16384,0.2798951201968723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,2560,0.05390222205056084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,2560,7168,0.01848088867134518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,768,8192,0.009194666312800514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,256,0.2727128929562039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,128,65536,0.2713182237413195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,256,512,0.009584888815879822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,51200,4096,0.9889013502332898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,256,128,0.0033413333197434745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,4096,256,0.006220444622966979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,4096,0.030050665140151978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,16384,0.023333332604832117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,51200,3584,0.05381066600481669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,768,2560,0.04050933321317037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,2560,0.02071644365787506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,3072,0.06057155794567532
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,7168,6144,0.019354666272799175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,512,3072,0.006640888750553131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,3584,2560,0.020392000675201416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,1024,65536,0.6597697999742296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,4096,3584,0.02922399838765462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,64,2048,0.0070088886552386815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,256,0.005396444350481033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,65536,0.15956978003184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,128,7168,0.043775998883777194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1536,1536,0.0063164445261160536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,51200,0.1458453337351481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,51200,2560,0.07720355855094062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,2560,6144,0.04446666770511203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2560,1536,0.024333332975705464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,12288,0.051926222112443715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,2048,0.006714666469229593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,128,0.003773333297835456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1536,6144,0.047351999415291675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,4096,128,0.014507555299335055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,3072,0.01034577770365609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,4096,0.010407110883129967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,8192,0.16993066999647352
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,7168,0.01827022267712487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,3072,0.007805333369308048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,256,0.005525333185990651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,64,64,0.005039999882380168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,128,10240,0.013544888959990608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,65536,3584,0.18290577994452584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,4096,768,0.005160888863934411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,256,2048,0.005863111052248213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,10240,12288,0.07520177629258898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,1024,51200,0.381056891547309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,16384,768,0.042809777789645724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,32,0.004780444420046276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,4096,512,0.0036924444139003754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,128,32,0.005203555441564984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,2048,0.009185777770148383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,65536,0.26439110438028973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,6144,0.006510222123728857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,4096,0.010956444674068026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,512,10240,0.016120889120631747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,8192,2560,0.12725244628058538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,768,3584,0.0070159998204973005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,5120,0.04111288984616598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,16384,0.018922666708628338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,1536,51200,0.04431644413206312
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,7168,16384,0.04902222090297275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,1024,7168,0.012463111016485425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,10240,65536,0.3656782309214274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,5120,1536,0.013312000367376538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2560,5120,0.09404889080259536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,3072,0.013993778162532382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2048,32,0.00574311117331187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3072,2560,0.06508088774151273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,10240,2560,0.020925333102544148
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,2048,0.052822222312291466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,6144,3584,0.02087822225358751
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,512,3584,0.11801332897610134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,7168,256,0.0200311111079322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,512,0.020863110820452373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,1024,0.007086222370465596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,32,5120,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,1024,0.8829706509908041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,8192,3072,0.10628444618648952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,6144,256,0.005743999862008625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,10240,16384,0.9020417531331381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,256,0.004380444271696938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,3584,128,0.0033208889265855155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3584,768,0.006592000110281839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,5120,2560,0.01147466649611791
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,6144,8192,0.038378665844599404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,5120,65536,2.0628159840901694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,16384,128,0.004874666531880696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,512,0.009927111367384592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,65536,1024,0.038366221719317965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,512,0.007278222176763747
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,256,12288,0.028307557106018066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,32,0.005689777847793367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3072,1024,0.006583111153708563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,3072,0.006590222318967183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,2048,0.030830221043692693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2560,8192,0.018615111708641052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,512,0.004450666821665234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,64,0.030444443225860596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,51200,0.07005333238177829
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,3584,0.0566435522503323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2048,256,0.002993777808215883
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,2048,0.023708444502618577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,128,128,0.003198222153716617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,32,64,0.0025946667624844448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,8192,0.2228622171613905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,65536,0.0921742253833347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,8192,0.016723554995324876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,64,0.007979555262459649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,2560,768,0.07657511366738214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,2560,256,0.0032293332947625052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,51200,16384,0.22971643341912162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,768,512,0.00868977771864997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,10240,0.1538426611158583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,512,0.025601777765485976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,16384,3072,0.02793333265516493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,64,7168,0.0063999998900625445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,8192,0.014502222339312235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,4096,0.06336355871624418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,7168,0.1260648833380805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,1536,0.013214222258991666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,32,768,0.005459555735190709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,10240,0.14257778061760798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,1024,1536,0.005516444643338521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,2560,0.01202933324707879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,2560,3072,0.024735111329290602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,256,2560,0.006031999985376994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,3072,0.02502044373088413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,10240,8192,0.030744890371958416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,3584,1024,0.005438222239414851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,2048,0.013189333180586496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,3584,0.1621075603697035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,256,0.004351110921965705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,256,3072,0.008001777860853408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,4096,51200,1.5682675043741863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,3584,0.0119991112086508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,32,5120,0.009964444571071202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,12288,0.024067555864651997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,12288,4096,0.27034221755133736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,6144,3072,0.1603511174519857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,512,0.003677333394686381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,4096,0.13656177785661486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3072,2048,0.010759111079904767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,256,32,0.0025759999536805684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,10240,0.013831111292044321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,3584,0.22796445422702363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,1536,2560,0.05524444580078125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,7168,0.104020443227556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,2560,16384,0.024579554796218872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,3072,32,0.00830666638082928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,256,0.0030666666312350165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,512,2048,0.0074515557951397365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,7168,16384,0.7911795510186089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,2048,0.09190222289827134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,5120,768,0.04563466707865397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,4096,0.011504888534545898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,6144,12288,0.030231111579471167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,256,51200,0.05571999814775255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,8192,0.0916631089316474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,512,51200,0.05415199862586128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,8192,0.04166400101449754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,65536,0.1481742196612888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,12288,0.0901671118206448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,51200,6144,2.0556986067030167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,2560,0.04903911219702827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,1536,3072,0.00721155520942476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,8192,2560,0.03189244535234239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,1536,768,0.01424533294306861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,64,5120,0.006532444308201472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,768,8192,0.009301332963837517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,3584,0.030540443129009668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,4096,32,0.0034551111360390982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2048,3584,0.06136000156402588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,5120,0.37654932339986164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,1024,4096,0.010324444207880232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,16384,1536,0.02646844420168135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,256,6144,0.012034667034943899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,2048,0.010148444109492833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,256,5120,0.04336177640491062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,5120,16384,0.06153866979810926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,7168,128,0.022936888866954382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,768,768,0.07524977789984809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3072,4096,0.01424533294306861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,5120,0.07517511314815946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,10240,65536,0.28626754548814565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,6144,0.013271110753218332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,512,2048,0.005370666583379109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,5120,51200,0.11055911249584621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3072,12288,0.026915555198987324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,65536,5120,1.2930142084757488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,5120,0.03233066532346938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,12288,51200,0.731047100490994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,64,0.0031031111462248694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,6144,0.11913600232866074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1536,16384,0.20173422495524088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,3584,0.20593955781724718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,128,6144,0.006575111299753189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,4096,0.013398222625255585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,768,8192,0.01529600057337019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,2048,512,0.016550223032633465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,256,0.06867733266618517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,8192,0.01463733282354143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2048,4096,0.03951911131540934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,128,51200,0.055408888392978244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,512,2048,0.007125332951545715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,16384,0.04298222064971924
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,51200,0.018837332725524902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,4096,0.1658737791909112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,1536,65536,1.1603884167141385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,128,0.023579556081030104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,8192,0.028578665521409776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,16384,0.036342223485310875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,2048,0.19221688641442192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,4096,4096,0.031269331773122154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,64,0.007904888855086433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,12288,8192,0.06508800056245592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,32,0.004295999805132548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,256,32,0.002814222127199173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,256,3072,0.008560889297061497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,6144,0.055480887492497764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,10240,0.37802754508124453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,3072,7168,0.1730604436662462
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,6144,0.2825911045074463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,1024,0.01242844429281023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,1536,2048,0.016175111134847004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,6144,0.06129066811667549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,768,64,0.0027591110103660156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,10240,2560,0.13582311736212835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,6144,0.02013066742155287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,1024,768,0.021723555194007024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,128,0.0027822222974565295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,4096,2048,0.010113777385817634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,10240,0.5646853446960449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3072,128,0.01126488877667321
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,1024,8192,0.06969866487714979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,12288,0.04962222112549675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,12288,3072,0.058683554331461586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3584,1024,0.004664888812435998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,12288,0.05424177646636963
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,5120,0.006253333141406377
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,65536,6144,0.5594906806945801
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,512,0.0041537777417235905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,1536,8192,0.03639555639690823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2560,2048,0.040214220682779946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2560,768,0.0075493330756823225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,1536,0.008848889006508721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,128,4096,0.007994666695594788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,2560,0.020389333367347717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,12288,256,0.050127112203174166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,65536,0.6459573109944662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,1536,0.01314933349688848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,65536,10240,4.464800092909071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,768,0.009464888936943477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,2560,65536,0.566673755645752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,256,0.00304888892504904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,768,2560,0.008110221889283922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,512,8192,0.06959022416008843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,3072,65536,0.8731822437710233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,5120,65536,0.18601955307854545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,1536,0.0075191110372543335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,768,0.010704889065689512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,5120,7168,0.017115554875797696
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,768,0.00886577781703737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,512,2560,0.007541333635648091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,128,5120,0.07140800025728014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,6144,1536,0.008064000142945183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,6144,2560,0.011384000380833944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,10240,1.4254897435506184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,16384,0.006520000182920032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,8192,0.1181173324584961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3072,10240,0.021469333105617102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,8192,10240,0.2907866636912028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,10240,0.02165777815712823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,1024,0.0041831110914548235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,512,0.04031199879116482
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2560,4096,0.020803555846214294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,3584,0.07957777712080213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,1536,5120,0.1197048823038737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,2048,0.01459111107720269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,32,0.008653333617581261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,6144,0.030418667528364394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,8192,0.0476435555352105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,64,128,0.005024000174469418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,4096,0.01757866640885671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,4096,5120,0.10372889041900635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,512,8192,0.09499466419219971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,768,12288,0.010434666441546546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,65536,51200,1.9901065826416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,512,1536,0.006309333360857434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,768,0.0038000001675552795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,8192,2560,0.012262221839692859
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,16384,0.03994399971432156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,2048,0.020555555820465088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,10240,256,0.0074924445814556545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,32,32,0.005037333402368757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,64,0.0025537777692079544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,5120,0.16809333695305717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,2560,0.06560889217588636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,2048,0.019088000059127808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,32,5120,0.009313777916961247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,256,12288,0.032060444355010986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,10240,12288,1.0151004791259766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,8192,0.2484204504224989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,8192,128,0.015271999769740634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,1536,0.00700711127784517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,10240,256,0.01258488910065757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,2048,128,0.006953777538405524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,5120,0.6344079971313477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,4096,0.035092443227767944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,32,3584,0.009304888546466827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,512,0.023149333066410486
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,512,6144,0.07766489187876384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,512,0.016760889026853774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,768,65536,0.07723377810584174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,128,1024,0.005809777726729711
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,7168,128,0.026147555973794725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,7168,12288,0.04756977823045519
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,3072,2560,0.013255111045307584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,5120,6144,0.16860444015926787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,65536,4096,1.3688142564561632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,4096,4096,0.011689778003427716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,6144,0.018358222312397428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,51200,0.05381244421005249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,7168,64,0.019401777121755812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,51200,768,0.051651553975211255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,1536,3072,0.009312888814343346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,2560,0.01202933324707879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,2048,0.005646222167544895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,3584,0.024200000696712073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,2048,0.3544150988260905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1536,3072,0.00815288888083564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,3584,0.03649866580963135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,3584,16384,0.31272000736660427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,51200,768,0.2810800075531006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,65536,768,0.505883534749349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,5120,0.029645333687464397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,7168,768,0.00814400033818351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,3072,0.017900443739361234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3584,8192,0.02307022280163235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,64,256,0.006171555568774541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,65536,2048,0.0432879991001553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,256,65536,0.07735821935865614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,512,65536,0.7349839740329318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,64,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,7168,1024,0.0891466670566135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,8192,32,0.035051554441452026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,768,7168,0.011185777684052786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,4096,256,0.020111110475328233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,16384,0.0064275554484791225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,256,256,0.0042151109211974675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,65536,0.2511306603749593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,64,3072,0.0063786668082078295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,51200,0.19939733876122367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,7168,0.02567200031545427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,10240,0.45426045523749453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,5120,0.015600000818570455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,2560,0.007644444704055786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,5120,0.17331733968522814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,2048,32,0.012123555772834353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,3072,5120,0.08833066622416179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,6144,10240,0.37361688084072536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,10240,0.015015999476114908
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,10240,0.037010666396882795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,65536,0.1946666638056437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,8192,16384,0.07474578089184232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,128,0.0033644443998734155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,6144,3072,0.014934221903483072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,1536,0.13769066333770752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,12288,0.029213332467608984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,16384,0.22108533647325304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,6144,0.16198400656382242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,512,5120,0.009476444787449306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,256,0.013562666873137156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,16384,256,0.00452977791428566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,6144,1024,0.013747555514176687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,768,0.010512000156773461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,1536,1024,0.007220444579919179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,7168,0.017996443642510306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,10240,0.42784889539082843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,256,0.005399111244413588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,3584,1024,0.007076444725195567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,2560,5120,0.10305866930219863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,5120,1024,0.007788444558779399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,5120,0.5270791053771973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,768,0.2975368764665392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,768,2560,0.031515555249320135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,10240,768,0.06655110915501912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,512,4096,0.027110222313139174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,12288,2560,0.019893333315849304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,1536,0.006732444382376141
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,768,16384,0.02127555509408315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,2560,128,0.009754666851626502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,10240,0.016543111867374845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,128,0.005113777601056629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,32,512,0.008833777573373582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,16384,64,0.03254755669169956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,2048,0.09664177894592285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,64,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,12288,4096,0.045927110645506114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,1024,7168,0.010285333626800114
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,128,1536,0.0058319999112023254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,4096,51200,0.6519422001308864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,256,3584,0.011032889286677042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2048,1024,0.0064897777305708984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,4096,0.1995368931028578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,8192,51200,5.109367794460721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,768,8192,0.049068444304996066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,65536,0.15505511230892605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,256,256,0.005708444449636672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,10240,768,0.010312889185216691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,32,65536,0.2935680018530952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,512,32,0.002704888789190186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,51200,0.3326115608215332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,768,0.07538577583101061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,5120,0.047925333182017006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,4096,0.011895111037625207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,1024,4096,0.011281777587201862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,7168,0.0181057784292433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,10240,5120,0.03285600079430474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,10240,4096,0.023431110713216994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,5120,10240,0.5269946522182888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,10240,51200,1.5806923972235785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,16384,0.24233955807156035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,16384,8192,0.04105688797103034
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,7168,0.01646577815214793
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,65536,0.08251466353734334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,16384,5120,0.6133022308349609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,64,3584,0.007460444337791867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,12288,0.12271111541324192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,4096,768,0.04593155450291104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1024,1536,0.01976355579164293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,2048,0.30460532506306964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,51200,2048,0.07821066512001885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,16384,4096,0.03442222211096022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,6144,7168,0.02034222251839108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,10240,0.1413875553343031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,1024,0.007130666739410824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,128,65536,0.011008888483047485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,8192,12288,0.036079999473359846
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,65536,10240,0.32595377498202854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,8192,768,0.007135999699433644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,256,64,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,12288,0.4924355612860785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,10240,1536,0.011167999770906238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,4096,2048,0.008255999949243333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,4096,0.2797022130754259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,3584,256,0.02243999971283807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,1536,0.008775110873911116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,512,0.018193778064515855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,768,6144,0.00996977753109402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,65536,128,0.009392889009581672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,256,8192,0.010964444114102257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,768,128,0.005160000175237656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,51200,3584,1.2974667019314237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,128,32,0.006648889018429651
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,12288,0.04936266607708401
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,2048,5120,0.01571822166442871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1536,65536,0.22515556547376844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,8192,65536,0.29546666145324707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,2560,0.008049777812427944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,128,3584,0.008007999923494127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,256,3584,0.008091555701361762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,512,3584,0.009192888935407003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,16384,0.18201777670118544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,6144,0.09359644518958198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,4096,64,0.0027359999302360746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3584,4096,0.031089779403474595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,1024,0.018319111731317308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,10240,0.03185866607560052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,65536,0.23057688607109916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,10240,3584,0.1345120006137424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,16384,0.049698667393790356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,10240,5120,0.11139378282758926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,12288,0.007598222129874759
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,3584,0.016582222448454965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,2560,5120,0.13244622283511692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,768,7168,0.010991111397743225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,3584,0.007672000262472365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,128,64,0.002970666728085942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,2048,51200,0.4306115574306912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,256,4096,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,768,2560,0.011415111521879831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,8192,256,0.008760000268618265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,3072,0.008276444342401292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,4096,1024,0.006184000107977126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2048,256,0.004611555486917496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,64,16384,0.029515554507573444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,128,0.0070159998204973005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,256,0.00700444479783376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,65536,2.0823947058783636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,512,0.006934222247865465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,64,1024,0.005103111267089844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,12288,512,0.012530666258600024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,5120,65536,0.10991022321912979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,3072,0.02442755632930332
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,16384,768,0.013366222381591797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,64,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,2560,2048,0.06491111384497748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,8192,32,0.005485333502292633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,6144,8192,0.06897422340181139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,3072,0.0951884455151028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,16384,3072,0.02124533388349745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,12288,0.04186844494607714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,51200,1.877780490451389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,7168,64,0.021867555048730638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,512,0.0063573333124319715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,16384,0.10291377703348796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,8192,12288,0.05787999762429131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,3584,0.013880000346236758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,51200,1.771569781833225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,3072,0.07344622082180448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,1024,6144,0.07488977909088135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,7168,0.37995288107130265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,65536,3072,1.3528728485107422
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,6144,1536,0.0693004462454054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,12288,512,0.03587111168437534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,3072,512,0.03395911057790121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,64,5120,0.007617777420414819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,4096,0.5268470976087782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,10240,0.0173333336909612
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,64,4096,0.011491555306646558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,3584,768,0.025814221964942083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,1536,12288,0.0277493331167433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,64,3584,0.00702044450574451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,768,3072,0.008039111064540016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,2048,12288,0.020833777056799993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,1024,64,0.002926222152180142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,65536,51200,17.541188557942707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,3584,0.009383111364311641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,128,0.008452444440788692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,10240,0.021693333983421326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,1024,0.005060444275538127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,3072,0.3069937758975559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,7168,4.54239018758138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,6144,6144,0.31167909834120006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,128,256,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,2560,1024,0.006007110906971826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,256,1024,0.005978666659858491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,2560,0.022384888596004907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,65536,0.2136684391233656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,5120,12288,0.03807555635770162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,16384,0.12562044461568198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,4096,8192,0.09416977564493816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,16384,6144,0.12866222858428955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,12288,7168,0.30521154403686523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,4096,0.1685075494978163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,3072,32,0.03721955418586731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,7168,10240,0.06222133504019844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,8192,0.008681777450773451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,6144,2048,0.043924444251590304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,65536,256,0.015751111838552687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,51200,0.13781244224972194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,12288,128,0.041033777925703264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,3584,3072,0.013812444276279874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,16384,0.10718844334284465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,3072,0.077674667040507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,3072,0.016154666741689045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,128,0.0028595555987623003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,8192,8192,0.03748355640305413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,3072,65536,0.14681866433885363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,1536,0.06922666894065009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,2048,5120,0.021740444832377966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,8192,1024,0.010027555955780877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,12288,8192,0.04545688960287306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,12288,3072,0.016915554801623028
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,512,65536,0.06891200277540419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,6144,51200,0.13701866732703313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,32,2560,0.019147555033365887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1024,7168,0.011090666883521609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,512,51200,0.0597626699341668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,256,768,0.008098666866620382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,65536,512,0.019910222954220243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,16384,12288,0.6582346492343479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,1536,0.014383999837769402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,32,65536,0.28733687930636936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,512,0.00757155567407608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,64,0.002565333412753211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,10240,32,0.025282666087150574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,10240,1536,0.0112631113992797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3584,4096,0.013463111387358772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,1536,2560,0.00740622232357661
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,2560,5120,0.5051564640469021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,12288,0.0328053335348765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,7168,4096,0.014840000205569796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,32,256,0.005439110928111606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,2048,512,0.018775999546051025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,1024,0.44739288753933376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,16384,0.02696177860101064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,10240,0.009176889227496253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,256,5120,0.00980711148844825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,7168,1024,0.008847110801272923
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,1536,0.010291555689440833
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,2560,3072,0.09455377525753444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,32,0.005836444596449534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,65536,51200,1.5034649107191298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,512,3584,0.02901155418819851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,10240,1536,0.013857777747843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,1536,3584,0.07320177555084229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,1024,0.006837333242098491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,1536,3072,0.00850755555762185
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,65536,0.1301528877682156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,512,768,0.005583111196756363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1536,128,0.005292444593376583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,1536,0.06940711206860013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,768,2560,0.02403733299838172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,7168,0.09216622511545818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,512,2048,0.049345778094397656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2048,65536,1.3508062362670898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,12288,8192,9.815850151909723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,3584,512,0.16144711441463894
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,51200,128,0.008316444853941599
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,51200,1536,0.050323555866877236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,768,65536,0.08444178104400635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,51200,65536,2.1488942040337458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,512,64,0.0041582224269707995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,6144,256,0.0063288890653186375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,6144,65536,3.646391126844618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,3584,0.07540444533030193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,10240,2560,0.22939732339647081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,65536,3072,0.9208836025661893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,6144,0.07744889126883613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,5120,512,0.0035617777870761026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,2048,4096,0.009338666995366415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,10240,10240,0.14596800009409586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,2560,256,0.005693333430422678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,1024,16384,0.34659822781880695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,65536,2.1821795569525824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,65536,51200,1.4955920113457575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,256,0.011327111058764987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,51200,768,0.047183109654320605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,2048,1536,0.02663999961482154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,12288,0.08284978071848552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,5120,0.030031998952229817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,128,2048,0.013758222262064615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,51200,3072,0.08047822448942396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,16384,3584,0.3855582343207465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,64,0.00543200017677413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,12288,0.049759109814961754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,7168,256,0.006105777704053455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,7168,256,0.022824888428052265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,8192,0.04685511191685995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,65536,10240,0.3543226718902588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,3072,5120,0.01463644372092353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,7168,0.03321511215633816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,65536,3.236318164401584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,12288,12288,0.0779635575082567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,5120,4096,0.16262400150299072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,10240,10240,0.27939645449320477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,32,0.01574488812022739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,16384,0.11402311589982773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,3072,3072,0.011720889144473605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,7168,1536,0.008771555291281806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,8192,768,0.006166222194830577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,8192,6144,0.027854220734702215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,51200,10240,3.214775933159722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,12288,10240,0.051148444414138794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,2560,12288,0.02316088808907403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,8192,3.7460505167643228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3072,1024,0.016723554995324876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,5120,0.05892355574501885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,10240,0.40321511692470974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,12288,768,0.05120088987880283
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,51200,6144,0.09275555610656738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,10240,6144,0.03232533401913113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,2048,256,0.005527110977305307
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,7168,8192,0.0262080000506507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1536,7168,0.016102222932709586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,2560,65536,1.3468906614515517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,16384,0.051479998562071055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,7168,2560,0.01941244469748603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,12288,2.411008834838867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,768,3072,0.008601778083377415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,128,0.0063991112013657885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,3584,0.06873244709438749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,512,0.028254220883051556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,7168,0.043776889642079674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,64,64,0.009076444639099969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,51200,8192,1.628037346733941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3584,51200,1.161995569864909
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,8192,64,0.0074373334646224976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,256,768,0.004641777939266629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,6144,0.06624799966812134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,8192,7168,0.02696888811058468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2048,3584,0.009469333622190688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2048,32,0.0058337777025169795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,4096,0.010611555642551847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,16384,0.137445330619812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,64,6144,0.010048000348938836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,4096,0.006464889066086874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,4096,1536,0.010677333507272931
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,6144,1024,0.027495111028353374
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,65536,2048,0.07599377632141113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,8192,5120,0.018491556247075398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,10240,1.3370960023668077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,1536,0.007841777470376756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,6144,0.01883555617597368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,6144,512,0.03334577878316244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,5120,0.09232799874411689
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,512,6144,0.04857777886920505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,12288,768,0.012139555480745105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,6144,3584,0.013047999805874295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,3584,0.08164888620376587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,5120,0.01680799987581041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,3584,12288,0.137445330619812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,65536,0.10956799983978271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,6144,0.11807466877831353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,32,0.002942222274012036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,128,512,0.0031911111954185697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,512,768,0.004859555512666702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,768,65536,0.06914044751061334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,1024,10240,0.10956711239284939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,2048,0.029800001117918227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,5120,768,0.03691555394066705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,256,768,0.00488711107108328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,10240,3584,0.19707022772894967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,768,16384,0.07897600200441149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,4096,6144,0.11029333538479275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1536,512,0.0058702222175068325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,512,5120,0.02308355602953169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,12288,0.17867110835181343
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,2048,0.08658577998479207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,51200,0.24947288301255968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,512,3584,0.008660444782839881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,2048,512,0.0058497778243488735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,128,0.003121777748068174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,512,1536,0.0047360000511010485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,1536,51200,0.0900497767660353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,12288,0.03270488977432251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,65536,768,0.019693333241674636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,32,10240,0.02502577834659153
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3072,1536,0.0075431110130416016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,512,4096,0.006305777778228124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,64,16384,0.018417777286635507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,4096,0.010769777827792697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,768,51200,3.0157032012939453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,4096,0.0177182224061754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,8192,512,0.03973866502443949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,3072,2560,0.021993777818149988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,768,128,0.0040995557275083335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,12288,1536,0.017103110750516255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,3584,768,0.005326222214433882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1024,3072,0.017607111069891188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,128,0.0598577790790134
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,7168,0.04193333453602261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,51200,4096,0.062457780043284096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,2048,2048,0.007333333293596904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,2560,4096,0.01292799992693795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,5120,10240,0.3192506631215413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,3072,0.047622223695119224
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,16384,0.041402667760849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,768,256,0.007291555404663086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,2048,0.013266666895813413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,4096,0.007894222107198503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,64,1536,0.004657777647177379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,7168,128,0.006817777951558431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,65536,8192,0.5766035715738932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,4096,5120,0.12323821915520562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,256,1536,0.006100444330109491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,51200,65536,1.753200848897298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3584,2048,0.01017599966790941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,128,12288,0.12011644575330947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,8192,0.41690222422281903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,7168,16384,0.3974471092224121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,12288,0.11853510803646511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,16384,2560,0.0328053335348765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2560,6144,0.013023999830087027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,51200,0.1694328917397393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,2048,0.6311422453986274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,16384,128,0.014988443917698331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,12288,65536,0.2523031234741211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,512,7168,0.05742488967047798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,2560,5120,0.012303999728626676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,12288,5120,0.02681511143843333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,5120,512,0.003735110991530948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,10240,16384,0.059968001312679715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,8192,5120,0.04698488778538174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,10240,6144,0.03275022241804335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,12288,3072,0.11634843879275852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,512,12288,0.012208888928095499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,65536,2048,0.8798053529527453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,64,10240,0.035769777165518865
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3072,128,0.004318222403526306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,10240,10240,0.0653217766020033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,7168,128,0.00627377794848548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,8192,8192,0.36351466178894043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,10240,0.05656444364123874
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,64,5120,0.009220444493823582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,768,256,0.008437333007653555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,768,10240,0.11420622136857773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,10240,51200,0.26855821079678005
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,65536,32,0.004493333399295807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,2560,10240,0.19600089391072592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,4096,2048,0.009090666969617208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,2560,2048,0.007035555938879649
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,32,6144,0.00870666652917862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,64,16384,0.038922667503356934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,8192,2560,0.022494221727053326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,64,12288,0.09407111008961995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,3584,5120,0.10744800170262654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,51200,3.801052517361111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,12288,16384,0.06427022483613756
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1536,6144,0.01642311116059621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,256,1536,0.013801777528391944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,16384,51200,0.4173297882080078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,51200,2048,0.3850568930308024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2560,1536,0.007638221813572778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,51200,0.062011553181542285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,16384,16384,1.6422559950086806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,64,16384,0.04587911234961616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,2560,0.00758755538198683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,1024,3584,0.010807111031479306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,4096,8192,0.023711999257405598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,128,2048,0.00813066628244188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,1536,768,0.004305777864323722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,6144,8192,0.08931466605928208
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,12288,32,0.00959911114639706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,1536,6144,0.10193244616190593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,10240,128,0.005122666557629903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,7168,0.0063457778758472866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,51200,10240,0.15166933006710476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,6144,128,0.021520889467663232
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,4096,128,0.01760088900725047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2560,16384,0.1430613332324558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,4096,0.021401777863502502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,8192,0.029883556895785864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,65536,1.7553279664781358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,5120,1024,0.009918221996890174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,7168,0.013637333280510373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,768,32,0.006488889041874144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,16384,1536,0.1676622231801351
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,768,128,0.004197333421972063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,5120,4096,0.22622044881184897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,51200,1024,0.11488533020019531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2048,4096,0.012457778056462606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2048,51200,1.0308728747897677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,128,0.003338666632771492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,5120,51200,0.7577395439147949
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,65536,2048,0.0753199987941318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,10240,128,0.006056889063782162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,3072,128,0.015615110596021017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,6144,2048,0.02702755563788944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,3584,2048,0.059938669204711914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,10240,128,0.005157333281305102
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,32,65536,0.07227466503779094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,16384,0.5236790974934896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,256,0.0052720002002186244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,10240,5120,0.022944000032213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,1536,0.0075795559419526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,65536,512,0.06902666886647542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,16384,7168,0.07542133331298828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2048,768,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,16384,256,0.005601777798599667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,4096,3072,0.01460977726512485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,3072,768,0.006228444476922353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,4096,2560,0.010086221827401055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,512,1536,0.02700888944996728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,64,8192,0.011721777419249216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,3072,0.05611822340223524
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,256,51200,0.05517333414819506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,16384,0.17455022864871553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,64,10240,0.06198933389451769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,2560,5120,0.06787111361821492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,7168,5120,0.02106488909986284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,256,2048,0.008926221893893348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,1024,0.017095110482639737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,128,0.040829333994123675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,7168,0.010986666712496014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,256,51200,0.052651554346084595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,8192,0.02972888946533203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3072,1024,0.00943288869327969
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,2048,8192,0.11315200063917373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,64,32,0.005186666631036335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,3584,0.006853333363930385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,4096,0.010977778169843884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,512,10240,0.008676444490750631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,65536,4096,0.13645422458648682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,51200,6144,1.423749287923177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,6144,12288,0.5135502285427517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,32,4096,0.02444800072246128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,128,0.029602666695912678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,256,0.00609688874748018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2048,12288,0.01995733380317688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,768,32,0.004025777594910728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,12288,16384,0.8620453410678439
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3072,512,0.031133333841959637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,7168,1024,0.011554666691356234
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,4096,1536,0.012520888613329994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,8192,4.2039031982421875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3072,1536,0.00777333312564426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,6144,512,0.02270577847957611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,256,0.005648889061477449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,128,32,0.0034479999707804788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,128,0.020618667205174763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,12288,2560,0.20179377661810982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,1024,64,0.008466666771305932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,10240,1024,0.008877333667543199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,256,0.009775111244784461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,6144,2560,0.01144444445768992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,3584,768,0.025437333517604407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,16384,0.42521601253085667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,8192,51200,5.047264099121094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,128,51200,0.06802933083640204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,512,16384,0.04311555624008179
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,16384,6144,0.06276800235112508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,3584,65536,0.0843857791688707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,16384,768,0.011174221833546957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,512,0.006034666465388403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,51200,7168,0.10294133424758911
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,10240,768,0.009121778110663096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,12288,0.03218577636612786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,7168,2.161581251356337
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,3584,10240,0.41912444432576496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,12288,12288,0.05376266770892673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,51200,6144,0.8738915655348035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,65536,0.9454995261298286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,32,0.002744888886809349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,4096,0.019002666076024372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,16384,768,0.011897777517636618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,128,1536,0.01643466618325975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,512,16384,0.8763795428805881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,10240,0.06925066974427965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,2048,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,65536,512,0.02511111067401038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,4096,0.007787555456161499
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,2560,12288,0.01854933301607768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,65536,10240,0.48316531711154515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3584,128,0.005921777751710679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,5120,1536,0.08732710944281684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,65536,0.45362313588460285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,6144,2560,0.014279999666743808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,51200,6144,2.0642221238878036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,8192,0.011104889214038849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,768,8192,0.01940444442960951
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,256,0.0470142232047187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,3072,12288,0.32612978087531197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2560,4096,0.0839395523071289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,768,32,0.0034231110993358824
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,256,0.007600888609886169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,12288,512,0.05248622099558512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,8192,0.011624000138706632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,2048,3584,0.060271998246510826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,1024,768,0.08018044630686443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,8192,2560,0.1475155618455675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,3072,768,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,64,2560,0.007667555577225155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,768,0.06231821907891167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,1024,0.05436888999409146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,51200,1.4184906217787001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,2048,12288,0.09777155849668716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,2560,512,0.005091555416584015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,1024,2560,0.00721955547730128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,64,2048,0.005295111073387994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,12288,8192,0.05370222197638618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,51200,6144,0.0942408906088935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,16384,1024,0.010788444843557147
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,2048,256,0.010752889017264048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,10240,8192,0.2681066724989149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,51200,0.24413156509399414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,128,256,0.0032702222880389956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,16384,0.26402399275037974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3584,3072,0.011365333365069495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,10240,8192,0.05670399798287285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,32,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,768,3584,0.009182222187519073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,10240,64,0.02159999973244137
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,7168,7168,0.027835554546780054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,2560,0.007701333198282454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,7168,0.0121742222044203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2560,65536,0.09460000197092693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,768,0.006405333264006509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,512,8192,0.042352000872294106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,512,512,0.009462222456932068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,16384,768,0.042578667402267456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1024,6144,0.012635555532243518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,768,2048,0.02604977786540985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,3584,0.010495111346244812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,6144,65536,0.2137288914786445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,3584,0.31100355254279244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,64,16384,0.10571466551886664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,16384,64,0.003002666764789157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,5120,0.021194666624069214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,3584,768,0.0053102220926019884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,5120,0.3424142201741536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,7168,8192,0.026556443836953905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,12288,10240,0.26112977663675946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,16384,14.802552964952257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,16384,3584,0.4863573180304633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,1536,0.01279022213485506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,65536,0.15278489059872097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,7168,1024,0.012310222619109683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,16384,1536,0.02530755599339803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,51200,0.5235715442233616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,16384,0.006924444602595435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,2048,768,0.005974222388532426
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,51200,12288,4.3547414143880205
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,64,8192,0.12186488840315078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,512,16384,0.02474755545457204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,12288,8192,0.0542791121535831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,2048,6144,0.12980355156792536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,2560,0.04591200086805555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,3584,5120,0.020021332634819877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,16384,2048,0.016020443704393175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,5120,12288,0.024393777052561443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,768,12288,0.08298044734530978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,5120,32,0.01722400055991279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,2560,3072,0.4478498035007053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,1024,16384,0.035007112556033664
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,2048,3584,0.021151110529899597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,128,768,0.007886221839321984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,3584,10240,0.026711111267407734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3584,3072,0.02492622203297085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,64,0.024477332830429077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,8192,12288,0.22747733857896593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,2048,0.0521617763572269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2560,2048,0.007383999725182851
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,12288,2048,0.015612444943851896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,8192,0.02114222281508976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,3072,12288,0.870058642493354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,256,0.005864000154866113
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,1024,8192,0.01735466718673706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,5120,0.008758222063382467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,2560,0.026524444421132404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,128,2048,0.007458666960398356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,1536,32,0.00267999991774559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,256,3584,0.010486221975750394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,10240,5120,0.03731733229425218
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,1024,6144,0.06976977984110515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,16384,768,0.009489778015348647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,768,65536,0.21039556132422554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,7168,7168,0.059417777591281466
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,5120,256,0.00582311095462905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2048,8192,0.03467466764979892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,3072,4096,0.03278755479388767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,1024,0.6289306746588813
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,5120,0.08420089218351577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,5120,0.256747563680013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,7168,3584,0.014908444550302295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,51200,6144,0.09316355652279323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,4096,12288,0.20671378241644967
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,1024,3584,0.008739555875460306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,3584,2560,0.01014311114947001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,5120,0.03831555445988973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,256,2048,0.024732444021436904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,2048,32,0.0027262222849660446
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,16384,1.6891697777642145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,64,12288,0.04099466734462314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,16384,0.05557688739564684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,16384,0.2894577715131971
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,512,6144,0.08347644408543904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,16384,0.03798133465978835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,256,32,0.005021333280536863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,64,2560,0.00675733346078131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,4096,2560,0.012235555383894177
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1024,768,0.011560888753996955
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,2048,32,0.007544000115659501
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,128,32,0.0027635554886526535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,6144,10240,0.026279111703236897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,768,0.00610133343272739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,7168,768,0.0057502223385704895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,4096,16384,0.030096888542175293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,3584,2560,0.08808177709579468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,768,1024,0.008646222452322641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,64,10240,0.00620977787507905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,1536,0.006824000014199152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,8192,256,0.02884977724817064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,1536,6144,0.04132444328731961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,3072,0.009309333231714036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,3584,51200,0.11921244197421604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,768,8192,0.06936888562308417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,2560,0.09182577662997776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,3072,2560,0.008262222011884054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,3584,0.007812444534566667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,8192,0.5397395557827419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,64,10240,0.008022222254011366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3584,12288,0.1275777816772461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,3072,0.006824888702895906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,51200,0.8408675723605685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,32,6144,0.011165333290894827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,51200,65536,0.9104248682657877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,16384,512,0.07561955849329631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,7168,1.3328746159871419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,7168,0.00887111077706019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,7168,12288,0.07677955759896173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,51200,7168,0.11204266548156738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3072,1536,0.012706666356987424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,1024,1024,0.006537777682145436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,64,12288,0.05660177601708306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,1024,0.007094221810499827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,5120,0.026751110951105755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,65536,8192,1.7333022223578558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,1536,0.011555555793974133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,64,12288,0.01566755606068505
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,3584,768,0.0037919998996787598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,3072,0.00758755538198683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,2560,0.02237866653336419
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,5120,7168,0.024138665861553613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,1024,0.008241777618726095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,7168,2048,0.015236443943447538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,3072,10240,0.30445154507954914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,16384,128,0.006792888873153263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,3072,1536,0.006786666810512543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,2560,12288,0.03159733282195197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,64,6144,0.006372444331645966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,6144,512,0.005212444398138259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,7168,0.02125155594613817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,3072,3584,0.05199377735455831
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,5120,10240,0.022203556365436975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,1536,65536,0.061343113581339516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,65536,1.0680373509724934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,10240,3584,0.021376889612939622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,64,4096,0.006055999961164262
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,5120,1024,0.02142133315404256
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,16384,3072,0.24466578165690103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,7168,2560,0.1166071097056071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,8192,8192,0.35520089997185605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,5120,512,0.005469333380460739
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,128,256,0.0075333333677715724
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,768,10240,0.08258755339516534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,128,0.0046773333516385825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,1536,0.007205333146784041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,5120,1024,0.0057964444988303715
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,1536,0.04031733340687222
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,6144,256,0.026188444760110643
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,256,0.00424177779091729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,128,128,0.0053137776752312975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2560,51200,0.15550577640533447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,32,512,0.005542222410440445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,32,0.004148444367779626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,32,1024,0.01683288812637329
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,768,2560,0.007909333540333642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,32,3584,0.007063111497296228
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,128,0.002744888886809349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,51200,2560,0.04551555713017782
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,8192,512,0.00573244442542394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,512,512,0.004458666675620609
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,3072,0.43500177065531415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,51200,8192,4.8891012403700085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,256,0.033321777979532875
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,51200,4096,0.06361422273847792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,5120,2048,0.01460800071557363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,5120,12288,0.04288088944223192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3584,16384,0.4285359912448459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,5120,0.02427822185887231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,65536,12288,0.39432266023423934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,6144,0.01332800007528729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,1024,2048,0.005702222386995952
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,7168,51200,0.21576533052656385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,2560,256,0.0030373332815037835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,8192,0.8332471317715116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,3072,51200,0.08673422204123603
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,4096,51200,0.13788178231981066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,10240,768,0.009121778110663096
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,6144,0.06566133101781209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,12288,1024,0.08065866761737399
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,3584,7168,0.09153688616222805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,1024,2048,0.02269777821169959
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,1024,0.810282654232449
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,51200,64,0.059487998485565186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,3072,2048,0.009124444590674506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,16384,0.03619022170702616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,5120,2560,0.019704000817404855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,51200,65536,1.9687813652886286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,4096,12288,1.6575440300835504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,10240,0.6233919991387261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1536,65536,0.8390933142768012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,10240,0.019129777948061626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,3072,128,0.01667022208372752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,768,12288,0.0714942216873169
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,1536,0.005000000198682149
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,3584,6144,0.16128533416324192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,3584,0.46096001731024844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,2048,6144,0.059782220257653125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,3072,65536,6.359881930881077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,128,0.004129777765936322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,2048,0.12287110752529568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,768,0.013032889200581444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,65536,1.5337368647257488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,10240,256,0.013654222091039022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,7168,2560,0.012246222131782107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,128,51200,0.07806666692097981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,10240,1536,0.011897777517636618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,2560,0.07413866784837511
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2048,256,0.011287111375066968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,10240,2560,0.01629688839117686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,1024,51200,0.3404124312930637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,6144,0.16194932990603977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,2560,0.2029199997584025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,256,7168,0.06441333558824328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,7168,512,0.017812444104088675
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,6144,7168,0.025394666526052687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,1024,0.0069715554515520734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,2048,8192,0.020206222931543987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,6144,256,0.020092444287406076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,1024,0.004027555386225383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,512,12288,0.06993599732716878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,10240,0.02365600069363912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,1024,2048,0.03576177689764235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,10240,256,0.03193155593342251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,128,3072,0.038635555240843035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,10240,64,0.00923644420173433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,512,5120,0.055587556627061635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,4096,4096,0.019534221953815885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,128,2048,0.005092444519201915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,51200,0.10958311292860244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,8192,256,0.005220444252093633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,512,32,0.005037333402368757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,3072,2560,0.07336533069610596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,65536,6.087980482313369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,3584,3072,0.0437466667758094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,5120,0.0636062224706014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,2560,7168,0.018402665853500366
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,256,4096,0.013517333401574029
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,1024,65536,0.2008755604426066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,2560,5120,0.02351377738846673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,7168,51200,0.11660088433159722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,16384,0.17464533117082384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,51200,0.33017778396606445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,3072,0.02961777647336324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,32,1536,0.005522666705979242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3584,256,0.02106933295726776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,32,0.002773333340883255
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,512,128,0.0057813334796163775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,128,0.004092444562249714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,768,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,128,0.0032239999208185407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,5120,3072,0.011197333534558615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,8192,7168,0.032794667614830866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,2048,0.005433777968088786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,256,768,0.005695999910434087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,6144,1536,0.009233777721722921
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,5120,10240,0.02872177627351549
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2048,5120,0.013080889152155982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,3072,32,0.006506666541099548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,7168,256,0.0063742221229606206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,256,0.21070310804578993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,65536,2048,0.6644444465637207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2048,10240,0.019056000643306308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,1536,0.026895110805829365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,6144,12288,0.5233795377943251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,7168,3072,0.01351111133893331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,256,0.015740444262822468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,12288,8192,0.8831324577331543
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,6144,2560,0.12003821796841091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,2560,8192,0.18267199728224012
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,51200,16384,14.754460652669271
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,4096,0.04717866579691569
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,6144,6144,0.02775555517938402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,10240,5120,0.30325243208143443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,7168,256,0.0032497776879204642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2560,4096,0.012038222617573209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,8192,0.12468088997734918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,5120,0.7104800012376574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,12288,51200,8.637186686197916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,16384,0.05818400118086073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,2560,4096,0.011159111228254108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3584,2048,0.01975555552376641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,8192,3072,0.10751911004384358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,3072,0.010928000013033548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,5120,2048,0.13561422295040554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,4096,65536,2.0906026628282333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,8192,256,0.004284444368547863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,512,16384,0.00979733301533593
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,4096,1536,0.03393777873780992
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,1536,0.06890577740139432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,7168,768,0.0058400001790788435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,2560,2560,0.011578666667143503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,256,0.0053004444473319584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,768,0.04307022359636095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,51200,3072,1.9370791117350261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,128,1024,0.004936888813972473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,2560,65536,0.11132178041670059
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,32,0.006640888750553131
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1536,6144,0.01761866609255473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,256,3072,0.029721776644388836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3072,10240,0.02313333253065745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,4096,0.08042577902475993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,6144,0.10710133446587457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,256,32,0.00526577772365676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1536,6144,0.08062844143973456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,5120,0.009887999958462184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,65536,0.07720355855094062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,3584,0.024551999237802293
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,10240,0.015442666080262927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,3072,2560,0.008822222550710043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,64,8192,0.006232888748248418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,10240,1.7389208475748699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,1536,10240,0.0887262225151062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,5120,128,0.007086222370465596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,10240,64,0.054005334774653115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,32,16384,0.008255110846625434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,3072,51200,1.301447974310981
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,3072,8192,0.05995644463433159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,1536,1024,0.0069822221994400024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,7168,0.157533327738444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,8192,768,0.006775110960006714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,1536,4096,0.03313688768280877
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,16384,0.07659288909700182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,12288,2048,0.0627688897980584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,12288,32,0.0035786665976047516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,32,10240,0.0068417779273457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,16384,512,0.013927110367351107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,512,65536,0.03238044513596429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,1536,768,0.014935111006100973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,51200,7168,3.6461779276529946
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,256,3584,0.032519999477598406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,10240,0.024775111012988623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,768,2560,0.009093333449628618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,6144,32,0.011672000090281168
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,6144,8192,0.13164710998535156
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,512,1536,0.008007999923494127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2048,8192,0.16806222332848442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,64,0.0026728889594475427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,7168,65536,2.756543901231554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,2048,64,0.006496888895829518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,10240,5120,0.6460924678378636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,6144,51200,0.18345155980851915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,1536,10240,0.05572888586256239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,256,4096,0.027092445227834914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,5120,0.07551199860042997
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,8192,5120,0.022991999983787537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,6144,51200,2.2409483591715493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,6144,65536,0.16008799605899388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,16384,128,0.03972355524698893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,5120,16384,0.04894666539298164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,7168,5120,0.01811822255452474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,256,0.003448888866437806
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,3072,256,0.00452533322903845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,51200,768,0.02347733411524031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,51200,4096,0.5355431238810221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,8192,0.017296888762050204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,3072,0.13670133219824895
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,16384,512,0.007655999726719326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,65536,1536,0.5877022213406032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,512,768,0.004942222187916438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,7168,0.02218311031659444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,1536,6144,0.019464888506465487
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3584,7168,0.09222844574186538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,1536,128,0.005364444520738389
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,256,0.003100444459252887
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,10240,128,0.005965333431959152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,1024,12288,0.17070666948954263
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,16384,1.5772231419881184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,3584,512,0.005607111172543631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,256,8192,0.015622221761279635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,2048,0.1600266695022583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,4096,0.08237422174877591
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,6144,2048,0.04190577732192146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,10240,4096,0.019489778412712943
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,10240,0.27571378813849556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,6144,3584,0.08192177613576253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,12288,12288,0.04968888892067803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,5120,512,0.017909333109855652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,10240,10240,0.0619751082526313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,16384,7168,1.7469235526190863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,128,2048,0.008039111064540016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,8192,0.02313333253065745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,768,2048,0.014287999934620328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,10240,0.5195955700344509
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,5120,2560,0.011868444581826528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,16384,0.02693155573474036
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,16384,512,0.050626665353775024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,1024,768,0.01424088908566369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,3072,0.006623999940024481
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,65536,0.2149440050125122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,12288,0.03562577896647983
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1024,128,0.0029031110720502008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,6144,2560,0.011129777464601727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,6144,1.0338426166110568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,65536,8192,1.1882088979085286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,128,0.013487999637921652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,5120,128,0.007224000162548489
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,4096,65536,0.1419395605723063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,1536,0.11205244064331055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,3584,5120,0.013307555682129331
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,51200,4096,0.06342311037911309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,2048,8192,0.05988978015051948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1536,4096,0.010022222167915767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,5120,4096,0.037978665696250066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,1024,0.02553066611289978
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,3072,51200,0.519627571105957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,16384,512,0.0099653336736891
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,5120,7168,0.11439377731747097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,8192,4096,0.017456889152526855
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,65536,0.19476532936096191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,7168,6144,0.23489154709710014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1024,256,0.009649777577983009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,32,128,0.005325333525737126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,768,256,0.004671111288997862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,16384,1024,0.08887110816107856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,2560,768,0.0038382220599386427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,3584,256,0.004691555682155821
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,2048,6144,0.018424888451894123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,8192,0.018538667096032035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,2560,3584,0.10406577587127686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,768,512,0.017483555608325534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,128,0.025303999582926433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,7168,10240,0.371413336859809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,3584,65536,0.08152621984481812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,5120,51200,0.13563022348615858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,2560,16384,0.06135288874308268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,12288,0.23716534508599174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,4096,2048,0.040272000763151385
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,32,6144,0.016125332978036668
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,6144,2048,0.01127466642194324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,32,512,0.007063999772071838
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,4096,65536,0.3761102093590631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,1536,0.0827253328429328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,5120,6144,0.3031200038062202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,10240,0.20417422718471953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,5120,128,0.028661333852344092
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,1536,0.02981511089536879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,768,12288,0.042034665743509926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,2048,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,32,10240,0.06616533464855619
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,2560,128,0.0033093332830402586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,2048,128,0.012124444047609964
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,65536,512,0.03292977809906006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,2560,2048,0.006820444431569841
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,256,8192,0.011770666473441653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,4096,1024,0.04153600003984239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1536,6144,0.013332444760534497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,16384,65536,12.9419674343533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,6144,0.12009599473741318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,1024,16384,0.02846933404604594
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,2560,1024,0.0181057784292433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,6144,0.013325333595275879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,2560,32,0.0028417776856157514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,2048,0.03878044419818454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2560,10240,0.2813928922017415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,7168,8192,0.02581688927279578
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,4096,2560,0.12975200017293295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,64,3584,0.0069395556218094295
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,256,8192,0.013532444834709167
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,128,1536,0.004639111045334074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,65536,6144,0.11825156211853027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,10240,65536,0.227659543355306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,5120,1024,0.017996443642510306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,3584,0.014289778139856128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,8192,12288,0.05681777662701077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,768,16384,0.021095999413066443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,3072,768,0.008467555873923832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,1024,8192,0.13003822167714438
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,2560,0.012992888689041138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,12288,2048,0.01713511182202233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,3072,51200,0.10845422082477146
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,256,16384,0.007958222594526079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,3584,128,0.005233777893914117
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,4096,2048,0.039206223355399236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,51200,512,0.017143110434214275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,16384,1536,0.021937777598698933
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,2560,0.013549333645237817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,12288,32,0.047714667187796704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,256,1536,0.006196444647179709
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,6144,256,0.007435555259386699
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,32,0.002748444469438659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,1024,3072,0.02408711115519206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,256,1024,0.004088888979620404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,3584,768,0.006456888798210356
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2560,7168,0.17367111312018502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16384,3584,12288,1.5174844529893663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,2048,0.04778844449255201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,7168,65536,0.30991554260253906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,1536,6144,0.01260444439119763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,10240,32,0.0075706665714581805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,1536,0.14170666535695395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,3072,256,0.019698666201697457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,97,2560,51200,0.09336266915003459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,2560,64,0.011786667009194693
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,2048,65536,0.06926400131649442
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1024,2048,0.007835555407736037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,51200,2560,0.1161315573586358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2560,32,0.006527110934257507
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,6144,0.01088177743885252
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,512,65536,0.022473777333895367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,3584,128,0.004259555704063839
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,3072,512,0.009675555759006077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,10240,16384,2.263696034749349
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,4096,3072,0.01699733402993944
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,4096,12288,0.2721911006503635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,3072,0.010446222292052375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,128,6144,0.01239911135700014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,65536,1.1774604585435655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,32,8192,0.012296000288592445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,16384,3584,0.22045332855648467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,128,10240,0.012067555553383298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,1024,3584,0.014861333701345654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,10240,768,0.010623999767833285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,12288,128,0.006130666782458623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,32,128,0.00508711114525795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,51200,0.36496178309122723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,7168,0.664555549621582
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3584,256,0.005688000056478713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,1024,7168,0.013884444203641681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,256,768,0.006039999839332368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,8192,768,0.06060444646411472
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,65536,5120,0.1689066621992323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,6144,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,1536,1536,0.005084444251325396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,6144,1024,0.0063688887490166565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,7168,0.04071822100215488
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,1024,6144,0.008223111430803934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,12288,65536,1.003047095404731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,4096,0.17967377768622506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,1536,65536,0.49687555101182723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,12288,0.01980177727010515
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,256,512,0.006471111128727595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,256,51200,0.07908888657887776
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,8192,0.08196177747514513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,256,65536,0.6812017758687338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,8192,7168,5.903885311550564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1536,12288,0.01772533357143402
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,512,12288,0.024127999941507976
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3072,256,0.004201777693298128
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,7168,3584,0.11358577675289577
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,512,512,0.003690666622585721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,51200,0.8238604333665637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,16384,6144,0.8210390938652886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,128,32,0.002630222174856398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,256,4096,0.010267555713653564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,5120,1536,0.04281422164705065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,6144,5120,0.019527110788557265
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,128,12288,0.021370665894614324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,6144,16384,0.06233511368433634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,1024,1024,0.005865777946180767
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,256,4096,0.00907555553648207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,65536,2560,0.6169573465983073
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,512,6144,0.012330667012267642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,7168,0.00834222220712238
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,768,768,0.016932444439993966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,10240,10240,0.060963557826148145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,8192,3584,0.12609955999586317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,7168,2048,0.011615999870830111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,128,16384,0.11454221937391494
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,512,3072,0.008034666379292807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,128,0.00526577772365676
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,64,3584,0.010306666294733683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,12288,512,0.007826666865083907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,1536,0.03250133328967624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,7168,1024,0.3474871052636041
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,128,16384,0.021536888347731695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,512,64,0.003030222323205736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,10240,3584,0.19082489278581408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,12288,6144,0.02749866743882497
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,6144,0.012268444730175866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,4096,32,0.0053280000057485365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,3584,2048,0.008475555313958062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,3584,0.07327822181913587
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,2048,0.008615111311276754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,1536,16384,0.20413245095147026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,3072,0.1035075585047404
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1536,51200,0.2696399953630236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,4096,0.08947377734714085
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,1024,768,0.00612088872326745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,768,0.010344888601038191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,3072,1536,0.031333333916134305
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,64,2048,0.021388444635603163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,7168,0.006155555446942647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,2560,0.012112889024946423
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,51200,10240,0.1493591070175171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,256,3072,0.006718222051858902
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,2048,2048,0.008087111016114553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,768,512,0.005410666680998272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,1024,512,0.0038773334688610504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,768,1024,0.004156444635656145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,7168,65536,2.990224838256836
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,10240,5120,0.35642398728264707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,1536,2560,0.008959111240175035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,6144,1024,0.009180444810125563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,51200,51200,0.7576630910237631
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,10240,0.0853911108440823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,12288,8192,0.05859466393788656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,5120,256,0.003563555578390757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,6144,0.014196443888876172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,16384,256,0.043649779425727
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,65536,64,0.04126222266091241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,1024,0.007794666621420119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,512,6144,0.0687555538283454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,4096,768,0.039911111195882164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,5120,0.012916444076432122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,3072,10240,0.35462133089701336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,8192,51200,1.169141345553928
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,512,2560,0.0068488890926043195
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,4096,1024,0.008382221890820397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,2048,6144,0.01240355521440506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,4096,0.033033778270085655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,2560,6144,0.015284443895022074
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,65536,2560,0.0665608909395006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,7168,0.023734221855799358
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,1536,7168,0.03556088937653436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,65536,7168,3.556060367160373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,65536,0.07211022244559394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,3584,0.016560888952679105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,512,1024,0.0047608887155850725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,65536,32,0.2699315547943115
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,10240,3072,0.0384568903181288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,5120,512,0.010850666297806634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,16384,256,0.00630044440428416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,4096,3584,0.129077328575982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,1536,0.011902222202883827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,64,2560,0.007115555306275685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,256,7168,0.013688888814714221
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,10240,4096,0.016664889123704698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,65536,12288,1.8406329684787328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,4096,12288,0.34563112258911133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,12288,0.06958222389221191
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,4096,64,0.003481777591837777
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,512,512,0.004886222382386525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,12288,256,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,512,2560,0.015544000599119397
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2048,8192,0.014084445105658637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,4096,0.037492444117863975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,12288,0.03131822082731459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,3072,32,0.005802666561471091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,3072,256,0.00544888898730278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,32,768,0.02595022155178918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,5120,16384,0.04106311003367106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,8192,65536,0.16739111476474336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,768,65536,0.08111644453472562
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,4096,0.022980444961124
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,768,4096,0.04319999946488274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,16384,5120,8.265061272515192
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3584,6144,0.2955226633283827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,65536,128,0.14393777317470974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,3584,0.01033155537313885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,51200,0.058719999260372586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,3072,5120,0.01497244503762987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,1536,0.026361778378486633
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,3584,5120,0.0191928893327713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,3584,0.0433991120921241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,4096,16384,0.0404151115152571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,128,0.0035013332962989807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,65536,768,0.04075288772583008
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,256,10240,0.015984889533784654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,8192,0.23674310578240287
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,1536,3072,0.06197066439522637
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,2048,128,0.0031386667655573953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,10240,16384,0.09033244185977513
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,10240,7168,0.27573778894212514
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,16384,512,0.007727111379305522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,12288,51200,3.0103128221299915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,64,12288,0.10338044166564941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,12288,2048,0.02494755552874671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,1024,8192,0.039566222164365984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1536,51200,0.0477164453930325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,2048,0.0058559998869895935
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,65536,12288,0.2735573450724284
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,3072,3072,0.03130311104986403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,1024,3072,0.0361422234111362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,128,65536,0.6713706652323405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,256,7168,0.05949866771697998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,1536,51200,0.6072808901468912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,7168,2560,0.1611537800894843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,16384,32,0.0075937774446275495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,128,128,0.006530666516886816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,7168,8192,0.022312889496485393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,10240,51200,0.15991289085812038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,512,2560,0.008005333443482717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,32,256,0.007280888656775157
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,3584,3584,0.014104000396198697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,2048,3072,0.008202667037645975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,65536,3584,0.0646595557530721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,2048,0.04133955637613932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,5120,256,0.016823111308945548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,12288,0.2133022281858656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,65536,3584,2.3300577799479165
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,2560,768,0.008360889222886827
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3584,768,0.005014222115278244
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,12288,10240,0.33815821011861164
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,256,128,0.0029173334025674393
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,1536,10240,0.07320088810390897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,128,0.018872888551818002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,16384,128,0.054823110500971474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,64,32,0.007677333222495184
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,16384,1536,0.01426666643884447
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,65536,3584,0.06555022133721246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,10240,10240,0.037920888927247785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,8192,2560,0.015122666954994202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,10240,128,0.006637333167923822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,256,512,0.00867377801073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,65536,1024,0.3301413324144152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,3584,0.2552399900224474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,65536,16384,0.290828439924452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1536,768,0.01274400038851632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,65536,1024,0.2327911059061686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,51200,768,0.33705690171983504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,512,65536,0.37035732799106175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,5120,128,0.006016888966162999
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,5120,512,0.024375999967257183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,1536,0.0692302253511217
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,51200,5120,0.07894311348597209
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,64,3072,0.023936000135209825
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,3072,7168,0.15490755769941542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,51200,4096,0.10815022389094035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,3072,3072,0.010871999793582492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,6144,0.12727288405100504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,2048,3072,0.008039111064540016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,2560,256,0.005386666705211003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,7168,64,0.007054222126801808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,12288,512,0.011157333023018308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,3584,32,0.0027182222240500977
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,3072,512,0.01086400035354826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,64,0.006528000036875407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,8192,2560,0.015692444311247934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,3072,8192,0.02702666653527154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,12288,256,0.004283555679851108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,1536,6144,0.05051555567317539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,5120,10240,0.03270755542649163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,8192,10240,0.044864889648225575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,10240,768,0.007703999678293864
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,64,256,0.0029502221279674103
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,2560,2560,0.008007999923494127
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,2048,2048,0.01606399979856279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,768,0.009946666657924652
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,12288,256,0.04997155401441786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,12288,768,0.030392888519499037
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,5120,65536,3.392463896009657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,256,0.009588444398509132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,2560,10240,0.19776266151004365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,256,0.011648000114493899
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,128,2560,0.006255110932721033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,51200,65536,1.7884701622856989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,4096,4096,0.011101333631409539
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,5120,1024,0.006535111202134027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,10240,1024,0.009901333186361525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,12288,10240,0.8662000232272677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,3584,10240,0.036087112294303045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,64,2048,0.007770666645632849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,128,0.02805066770977444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,16384,3584,0.36078400082058376
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,256,1536,0.006007110906971826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,128,64,0.007445333732499017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1024,7168,0.02858577834235297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,7168,6144,0.027594667341974046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,32,65536,0.01275022245115704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1792,8192,8192,0.2413555516137017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,51200,2560,0.05009333292643229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,7168,3584,0.08710400263468425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,16384,1536,0.01572088897228241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,1024,65536,0.5753991338941786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,12288,512,0.2238719993167453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,64,0.003947555604908201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,7168,16384,0.03970844546953837
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,6144,65536,2.4864879184299045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,8192,5120,0.018533334136009216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,32,65536,0.008295111358165741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,16384,1024,0.16580711470709905
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,16384,128,0.007336888876226213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,6144,64,0.02731288969516754
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,6144,7168,0.172088000509474
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,51200,12288,1.31895997789171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,128,256,0.0058151111006736755
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,32,2560,0.009073778159088558
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,12288,6144,0.6714897685580783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,6144,1536,0.2448071108924018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,10240,64,0.034854223330815635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,2048,32,0.003593777616818746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,65536,2560,0.34043110741509336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,65536,0.32507112291124135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,2048,1536,0.023349333140585158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,4096,2560,0.037075555986828275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,2048,32,0.009880888793203566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,12288,768,0.0451217790444692
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,7168,2560,0.01349511080318027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,32,64,0.002565333412753211
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,5120,16384,0.04771911104520162
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,2560,6144,0.013117333253224691
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,768,0.0166666673289405
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,3584,2560,0.00867377801073922
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,128,1024,0.0058746664888328975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,768,2048,0.012880889077981314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,5120,5120,0.1197742223739624
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,768,2048,0.019499555230140686
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,256,0.004956444518433677
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,51200,0.1641413370768229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,4096,768,0.005942222144868638
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,2048,12288,0.024871110916137695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,6144,12288,0.04037955403327942
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,4096,0.08709333340326945
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,3584,7168,0.18932888242933485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,3584,3072,0.0564240018526713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,8192,2560,0.013553777502642738
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2560,3072,0.07190310955047607
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,16384,32,0.026211554805437725
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,4096,3584,0.04292088747024536
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,8192,51200,5.2325316535101996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,10240,0.027092445227834914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,3584,3584,0.01333866682317522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,768,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,65536,6144,1.8251654307047527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,256,12288,0.021998221675554912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,2048,4096,0.014743111199802823
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,256,6144,0.009287999735938178
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,12288,65536,2.3861083984375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,32,5120,0.012821333275900947
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,12288,12288,0.08733599715762669
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,4096,3072,0.011133333047231039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,4096,7168,0.029512001408471003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,256,65536,0.017893332574102614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,12288,0.010341333018408881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,65536,10240,0.8107937706841363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,5120,2048,0.033803555700514056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,4096,12288,0.030994666947258845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,512,768,0.006610666712125142
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,8192,32,0.02466222147146861
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,8192,4096,0.0680968893898858
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,32,0.007695111135641734
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,6144,768,0.0269804447889328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,65536,512,0.027013333307372198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,16384,512,0.007798222204049428
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,6144,2560,0.011018666956159802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,65536,3584,0.06979644298553467
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,1024,7168,0.01372977760103014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,7168,768,0.006103111224042044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,51200,1536,0.41435911920335555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,512,12288,0.01921955578856998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,7168,1536,0.010113777385817634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,768,7168,0.10508888959884644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,3072,0.008546666966544257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,12288,16384,0.10492710934744941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,65536,2048,0.06719466712739733
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,12288,512,0.03361777795685662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,10240,768,0.06361599763234456
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,81,2048,65536,0.07435288694169787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,32,10240,0.4999190966288249
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,128,0.003129777808984121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,8192,3072,1.4924195607503254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,3072,7168,0.0708364446957906
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,1536,65536,3.8170979817708335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,8192,0.0288355549176534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,3072,0.01666844387849172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,256,8192,0.0210524449745814
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,512,8192,0.0384133325682746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,6144,4096,0.12534222337934706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,128,0.016859556237856548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,512,4096,0.008922666311264038
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,256,0.016789333687888253
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,10240,3584,0.06614933411280315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,4096,64,0.0058133333093590206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,768,12288,0.21032888359493682
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,7168,3584,0.05628089110056559
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,5120,0.023443554838498432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,12288,3584,0.025105777713987563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,2560,32,0.002642666714058982
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,1024,5120,0.009584888815879822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,6144,0.013606222139464485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,51200,0.10896178086598714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,51200,5120,0.1341244379679362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,5120,128,0.014259555273585849
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,7168,12288,0.07802133427725898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,7168,128,0.024209777514139812
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2560,8192,0.09926133023367988
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,4096,0.008171555896600088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,128,51200,0.057767113049825035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,10240,10240,0.03683288892110189
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,512,2048,0.007299555672539606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,1536,0.006330666856633292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,4096,12288,0.023994666006830003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,3584,0.008731555607583787
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,12288,51200,0.33022578557332355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,128,32,0.0027253333893087176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,4096,512,0.005575110928879843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,16384,12288,0.11114666197035049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,4096,128,0.004322666674852371
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,10240,0.1305991146299574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,512,3584,0.009091555244392818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,768,0.013444444371594323
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,128,2048,0.005071111023426056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,12288,12288,0.05122933453983731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,65536,1024,0.02334755493534936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,768,2048,0.011177777416176267
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,2048,4096,0.01072088877360026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,16384,128,0.04456177684995863
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,51200,3072,0.07889155546824138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,10240,0.03726222117741903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,10240,2048,0.9897680282592773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,6144,0.0298942228158315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,12288,256,0.0036657779581016968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,2560,3584,0.010627555350462595
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,7168,0.15522755516899958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,2560,3584,0.010631110933091907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,3072,12288,0.024336889386177063
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,16384,128,0.004964444372389051
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,12288,6144,0.050428443484836154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,2560,0.04503200120396084
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,64,8192,0.07052444749408297
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,6144,6144,0.02255200015174018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,6144,0.11548533704545762
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,2048,32,0.005050666630268097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,1024,51200,0.053299556175867714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,768,768,0.018477333916558158
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,64,768,0.0037324445115195382
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,6144,0.009929777847396003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,7168,256,0.018760888112915885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,10240,64,0.0028355556229750314
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,7168,768,0.0057946667075157166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,256,7168,0.011012444065676795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,7168,5120,0.3795759942796495
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,65536,32,0.16993599467807344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,1536,51200,0.05825689103868273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,51200,128,0.03037422233157688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,65536,256,0.02469066613250309
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,65536,128,0.014091556270917257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,7168,10240,0.026705778307384912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,64,2048,0.005621333503060871
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,4096,32,0.025337777203983728
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,2560,12288,0.023387556274731953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,2048,6144,0.011788444386588203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,256,3072,0.008609777523411645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,1024,3072,0.043191111750072904
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,64,0.0070062221752272705
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,32,6144,0.010078222387366826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,3584,0.007893333004580604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,7168,1536,0.00872444444232517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,12288,256,0.0064275554484791225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,16384,4096,0.2928782304128011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,768,768,0.0151591118839052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,2560,64,0.010842666857772403
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,12288,2560,0.016581333345837064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,2048,32,0.005495999836259418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,128,256,0.007052444749408298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,12288,6144,0.025570667452282373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,32,3072,0.010798222488827176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,7168,16384,0.04955733484692044
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,3584,16384,0.027913779020309448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,16384,2048,0.18197955025566948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,128,6144,0.03214044372240702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,2048,3072,0.05511377917395698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,1536,12288,0.017794667018784415
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,64,1536,0.03547199898295932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2305,12288,16384,0.8233431180318197
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,1024,128,0.008028444316652086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,32,5120,0.023334221707450017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,4096,51200,0.07004711363050672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,256,3584,0.03296444482273526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,768,3072,0.026573333475324843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,2560,3584,0.009834667046864828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,2560,0.05080355538262261
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,4096,10240,0.028628445333904688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,2560,64,0.0034622223012977173
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,3072,6144,0.10282666815651788
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,512,65536,0.06841599941253662
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,128,768,0.004504888835880491
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,5120,0.13743555545806885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,12288,3584,0.019669334093729656
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,256,3584,0.017148445049921673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,10240,512,0.035354667239718966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,3584,7168,0.023386667172114056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,10240,3072,0.01831733352608151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,3072,0.03238222334120009
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,1536,0.009764444496896531
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,16384,51200,0.24678132269117567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,5120,1024,0.013528000149461957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,5120,8192,0.09028266535864936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1536,12288,0.017471111483044095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,7168,64,0.01919822229279412
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,4096,7168,0.06120710902743869
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,2048,0.011213333242469363
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,32,4096,0.0090515555606948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,51200,2560,0.12331288390689427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,4096,2048,0.052271998590893216
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,6144,12288,0.04806133442454868
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,65536,1536,0.5067209137810601
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,5120,0.03741600116093954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,1024,5120,0.01275466630856196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,768,5120,0.024667556087176006
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,128,0.019908444748984445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,1536,10240,0.02388444377316369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,10240,10240,0.03344355689154731
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,512,256,0.002895111011134254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,512,256,0.003154666680428717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,2048,0.10977777507570054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,7168,10240,0.0369102226363288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,768,51200,0.025008888708220586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,3584,0.27164443333943683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,512,0.0037644443412621817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,256,16384,0.10777688688702053
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,3072,3072,0.01370133293999566
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,512,2048,0.007818666597207388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,10240,2560,0.01477866702609592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,10240,128,0.005033777819739448
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,8192,64,0.006583999842405319
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,225,6144,7168,0.030930668115615845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,768,512,0.01645599967903561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,3072,0.00944088896115621
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,256,12288,0.01973600023322635
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,10240,3584,0.15488621923658583
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,51200,6144,3.6610844930013022
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,256,256,0.005439110928111606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,512,0.05911022424697876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,17,128,512,0.005479111025730769
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,64,4096,4096,0.013473778135246701
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,3072,2560,0.01055466632048289
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,768,1536,0.005949333310127258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,1536,2560,0.007995555798212687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,51200,3072,0.8960373136732313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,51200,0.2558373345269097
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,4096,2560,0.013974222871992322
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,3072,0.014187556174066333
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1536,5120,0.033722665574815534
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,6144,7168,0.0707884430885315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,512,16384,0.022175111704402502
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3072,8192,0.026521777113278706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2560,16384,4.0915811326768665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,256,1024,0.020233778489960563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8,768,512,0.004326222257481681
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,7168,65536,0.14790311124589708
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,51200,256,0.010657777388890585
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,225,5120,3072,0.017119111286269296
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,6144,5120,0.024417777856190998
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,256,5120,0.031215998861524794
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,32,6144,0.054168889919916786
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,128,768,0.005192000004980299
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,16384,8192,0.06858399841520521
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,1024,512,0.003744889050722122
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,12288,65536,0.3366746637556288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,193,512,128,0.004287111262480418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,2048,0.17302310466766357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,1536,0.035198221604029335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,512,512,0.004423999951945411
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,1536,5120,0.0642355548010932
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,2048,2560,0.007161777880456712
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,768,1536,0.006843555718660355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,10240,4096,0.2331155671013726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,3584,6144,0.1840355528725518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,7168,12288,0.7183955510457357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1024,768,0.013557333085272046
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,256,3072,0.007536888950400882
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,1536,64,0.008852444589138031
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,8192,6144,0.07346755266189575
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,51200,32,0.09184533357620239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,3072,51200,0.9130480024549695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1281,4096,6144,0.10616444216834174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,3072,51200,0.08558311065038045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,4096,4096,0.143178661664327
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,768,0.009937778115272522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,8192,16384,0.07359200053744845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,6144,128,0.011961778004964193
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,32,12288,0.05666666560702854
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,12288,64,0.0029155556112527847
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,3072,5120,0.012896888785892062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2048,4096,0.036235554350747004
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,10240,64,0.006009777800904379
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,51200,8192,0.24112444453769258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,51200,0.17138132784101698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,256,32,0.007725333174069722
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,10240,1536,0.01329066687160068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2560,5120,0.01256444470749961
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3,5120,768,0.00657066661450598
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,7168,51200,2.226855172051324
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,512,3584,0.010071111222108206
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,65536,0.07938310835096571
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,32,7168,0.008179555336634317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3584,51200,0.10094133350584243
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,3584,0.05926577912436592
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,512,1024,0.004062222109900581
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,3584,2048,0.011511111425028907
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,3072,12288,0.019073777728610568
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,3584,65536,0.6698488659328885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,513,768,2048,0.012599111431174807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,12288,512,0.00797333319981893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,1536,6144,0.011787555283970304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,16384,8192,0.03945511248376634
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,1024,0.02309866746266683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,10240,0.01887022289964888
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,512,768,0.0037760001917680106
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,32,0.007057777709431118
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,65536,51200,1.7799563937717016
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,3072,10240,0.019630221856964957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,512,2560,0.020727111233605277
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,65536,0.09745955467224121
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1024,1024,0.006317333214812809
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,512,2048,0.006113777971929974
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1024,5120,0.03908888830078973
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,768,3072,0.018070222602950204
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,512,2048,0.007266666326257918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,65536,6144,1.3624827067057292
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,16384,5120,0.04756444361474779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,3072,5120,1.5554089016384547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,65536,2048,0.04614311125543383
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,32,128,0.005959110955397288
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,12288,1536,0.17652532789442274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,12288,0.18137778176201713
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,51200,0.18867644998762342
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,3584,16384,0.040431999497943454
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,8192,51200,2.703506681654188
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,256,3072,0.008776889079146916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,6144,0.006691555596060223
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,1536,6144,0.07243644528918795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,8192,32,0.0034382223255104492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,3584,16384,0.03420622150103251
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,512,65536,0.048735999398761325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,1024,2048,0.00657155571712388
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,32,512,0.003735110991530948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,1536,128,0.004359999878538979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,2048,6144,0.013789333403110504
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,64,0.006755555669466655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,4096,5120,0.2159359984927707
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,256,2048,0.007735110819339752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,6144,5120,0.41534490055508083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,16384,2048,0.2270533243815104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,2560,10240,0.13302844100528294
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,512,8192,0.03868888815244039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,2560,128,0.0058115555180443665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,3072,512,0.01663822266790602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,16384,768,0.009469333622190688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,32,0.007510222494602203
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,65536,51200,42.92407904730903
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,4096,128,0.0069875551594628235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,768,2048,0.006790222393141852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,12288,10240,0.589953793419732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,1024,64,0.003168888803985384
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,3584,2048,0.009167999857001835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,8192,5120,0.11327732933892144
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,8192,1536,0.011612444288200803
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,10240,65536,0.8565413157145182
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1536,7168,0.015466666883892484
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,4096,6144,0.015437333120240105
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,51200,256,0.014090667168299357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2816,7168,7168,0.22983198695712617
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,1024,256,0.007752888732486301
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,512,0.00563466673096021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,12288,0.22174399428897432
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,2048,128,0.0046302220887608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,64,8192,0.010676444404655032
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,256,3072,0.006710222197903528
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,4096,0.01792622274822659
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,1536,512,0.005003555367390315
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,160,4096,65536,0.23721867137485078
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,64,0.021685333715544805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,3072,768,0.01627200014061398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,12288,256,0.018002667360835604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,7168,65536,0.314830223719279
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,256,1024,0.004034666551484002
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,32,512,0.003318222239613533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,2048,12288,0.02013688948419359
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,128,10240,0.013463999662134381
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,2048,256,0.004560889055331548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,16384,8192,0.06411111354827881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2816,1024,10240,0.07901244693332247
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,128,0.005004444470008214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,65536,256,0.011337777806652917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,5120,1024,0.007364444434642792
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,11,256,12288,0.015832889411184523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,6144,32,0.007595555649863348
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,65536,256,2.0754826863606772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,10240,768,0.05828622314665052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,8192,51200,0.1508071157667372
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,12288,51200,4.003760867648654
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,5120,6144,0.07810311184989081
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,2048,1536,0.00720977783203125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,7168,128,0.010794666906197866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,10240,2048,0.01773599949147966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,512,0.01836266616980235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1025,1536,64,0.0037573331760035623
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1536,256,0.005569777968857024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,128,0.0033413333197434745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,65536,1.9944595760769313
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,16384,4096,0.025944888591766357
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,1536,1536,0.0079297779334916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,16384,0.3192959891425239
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,512,128,0.0028115556471877625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,16384,0.04427199893527561
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,768,1024,0.004632888982693355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,64,5120,0.012886222038004132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,5120,0.09115111165576512
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,128,0.0053022222386466126
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,32,10240,0.008741333252853816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,5120,3584,0.011982222398122152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,8192,8192,0.03928800092803107
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,5,51200,512,0.014339556296666464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,64,2560,0.00703999979628457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,64,32,0.003399111123548614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,81,128,768,0.004843555390834808
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,1024,51200,0.09881244765387641
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,256,2048,0.005413333161009683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,128,256,0.006247111078765657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,51200,7168,0.10411733388900757
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,2048,0.20234044392903647
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,4096,512,0.016993777619467843
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,6144,256,0.004781333522664176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,32,2048,0.007066667079925537
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,4096,0.14693689346313477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,768,3072,0.006671999891599019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,10240,3072,0.018516444497638278
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,128,6144,0.01240711079703437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,2048,0.060943113432990186
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,64,1536,0.0063235556913746735
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,10240,12288,0.18120177586873373
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,1024,2560,0.007083555890454187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,16384,8192,0.7554311222500272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,7168,16384,0.05877866347630819
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,5120,768,0.05539022220505608
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1536,5120,0.0841182205412123
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,33,768,51200,0.058119999037848584
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,12288,3584,2.368807050916884
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,16384,6144,0.06028978029886881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,51200,512,0.12613244851430258
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,8192,2048,128,0.021564443906148274
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,512,51200,0.06949688990910848
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,3584,3584,0.013591110706329346
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,12288,12288,0.6977395481533475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,9,8192,6144,0.025242666403452556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,128,3072,0.008650666309727563
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,2560,768,0.006280000011126201
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,1024,6144,0.02554399934079912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,512,12288,0.03439822130733066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,1536,8192,0.1381395525402493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,512,10240,0.12460800011952718
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3329,51200,1536,0.4227457841237386
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,8192,2048,0.014019555515713163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,51200,0.1724088854259915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,3072,1536,0.03557777735922072
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,256,0.05565777752134535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,5120,1024,0.03142755561404758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,256,8192,0.040567109982172646
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,7168,3072,0.06335200203789605
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,2560,7168,0.028159111738204956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,10240,7168,0.040780444939931236
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,12288,3072,0.23786489168802896
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,6144,1536,0.04826044374042087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,10240,8192,0.04822488956981235
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,3072,12288,0.09314844343397354
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,1024,1536,0.005530666559934616
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,5120,16384,0.049116445912255176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,6144,8192,0.022103110949198406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,128,65536,0.5752835803561741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,8192,0.06881244315041436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,512,1024,0.01125600023402108
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,2560,3584,0.013023999830087027
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,6144,65536,0.21315111054314506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,5120,16384,0.07666666640175714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,3072,4096,0.013658666776286231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,3072,0.007430222299363878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,1024,6144,0.08752799696392483
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2561,10240,512,0.03595111105177138
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3584,5120,12288,0.46427732043796116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,2560,2048,0.008969777988062965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,7168,12288,0.6806942092047797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,2560,128,0.011513777905040316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,12288,0.9434106614854602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3072,512,32,0.0037706668178240457
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,64,32,0.005589333259397083
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,15,2560,512,0.005039111193683413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,65536,3584,0.27523554695977104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,768,7168,0.008956444760163626
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,64,512,0.007302222152551015
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,65536,128,0.013651555611027611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,64,512,0.0034106667670938703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,3584,2560,0.00814577771557702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,12288,1024,0.08558311065038045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,6144,1536,0.01590755581855774
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,512,768,0.008781332936551835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,6144,0.35733866691589355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1025,2560,256,0.008405333591832055
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,65536,5120,4.448463016086154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,8192,16384,1.373851564195421
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,8192,512,0.009464888936943477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,10240,1024,0.07389511002434625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,51200,10240,0.1458639966117011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,128,32,0.002887111157178879
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,128,1536,0.018292443619834054
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,128,10240,0.03611733184920417
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,5120,6144,0.023543111152119104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,7168,1024,0.057130667898390025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,6144,2560,0.010934222075674268
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3841,10240,256,0.034134222401512995
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,256,0.00925688859489229
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,2048,0.031155556440353394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,256,5120,0.0068711112770769335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,51200,3584,0.15450310707092285
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,7,768,8192,0.012606222596433429
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,32,10240,0.006594666590293248
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,4096,12288,0.03611377875010172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,12288,768,0.009303111169073317
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,1536,0.019148444135983784
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,12288,6144,0.23986755477057564
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,12288,512,0.007100444700982835
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,3072,65536,1.1621821721394856
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,65536,1536,0.1477457814746433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3328,2560,65536,1.6361413531833227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,65536,51200,25.735114203559025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,2560,12288,0.15090489387512207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,1536,5120,0.014522666732470194
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,6144,0.01182044463025199
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3840,256,4096,0.017811555001470778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,256,768,0.007112888826264276
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1792,16384,4096,0.1721271144019233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,128,4096,0.008740444150235917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,3072,6144,0.015392000476519266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,6144,3584,0.08317955334981282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,128,8192,0.011900444825490316
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,65536,5120,0.0916737781630622
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,3072,768,0.018548443913459778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,128,768,0.0037502224246660867
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,128,5120,0.00628266649113761
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,10240,256,0.023299554983774822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,51200,1024,0.020784889658292133
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,6144,7168,4.305907355414496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,12288,7168,0.2630551126268175
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,51200,10240,1.495545811123318
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,256,256,0.004659555438492033
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,7168,10240,0.02846755584081014
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,8192,32,0.024805333879258897
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,51200,512,0.21374221642812094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,8192,65536,7.252091301812066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,128,16384,0.12929511070251465
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,64,6144,0.05113688773579068
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,1536,32,0.005647111270162795
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1280,1024,32,0.008560000194443597
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,8192,3584,0.014397333065668741
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,16384,512,0.009711110757456886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,4096,256,0.021366222037209406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,768,128,0.0028453332682450614
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,512,65536,0.06545333067576091
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16,1024,1024,0.003979555434650845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,768,51200,3.9944419860839844
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,3072,128,0.0036959999965296853
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,12288,3584,0.02741333345572154
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,51200,1536,0.051051557064056396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,12288,10240,0.6044995519849989
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1281,6144,512,0.023183999790085688
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,257,3584,6144,0.025088000628683303
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,4096,768,0.023297778434223596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,64,0.015124445160230001
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,513,3584,32,0.008740444150235917
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32768,128,256,0.031175110075208876
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,768,12288,0.009886222581068674
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,2560,6144,0.014198222094111972
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,3584,65536,1.5469129350450304
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,51200,51200,6.371936798095703
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,2048,5120,0.017154667112562392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,2560,256,0.021500445074505273
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,3584,3072,0.04475377665625679
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,15,2048,512,0.003297777846455574
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,65536,768,0.37701956431070965
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,48,51200,6144,0.08841244379679362
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,128,16384,0.2712106704711914
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,64,64,0.0074515557951397365
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2,10240,8192,0.032059556908077665
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,6144,1024,0.021558221843507554
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,128,6144,0.011885333392355176
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,512,8192,0.022122666239738464
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,65536,12288,0.25559377670288086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,32,3584,0.008029333419269985
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,2048,0.007157333195209503
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,2560,12288,0.2482231193118625
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,3584,512,0.01127022256453832
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,65536,12288,0.3263964388105604
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,12288,2048,0.04383288820584615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,768,3584,0.0075457774930530125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,7168,2048,0.011083555718262991
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,2048,16384,0.18693066967858207
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,65536,128,0.07720355855094062
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,80,10240,7168,0.0505617790751987
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,8192,128,0.0032702222880389956
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,5120,16384,0.15065599812401667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,16384,1536,0.01536622146765391
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,8192,3584,0.022008889251285132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,256,2048,0.019520888725916546
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,3584,12288,0.02052977846728431
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4353,512,4096,0.046837333175871104
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,51200,128,0.0070471109615431885
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2560,2048,0.0068853331936730286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,2048,1024,0.015185778339703878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,1024,65536,0.041456000672446355
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,3584,64,0.01572355628013611
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,768,3584,0.009391110804345872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,3072,0.04493244489034017
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,512,256,0.0052391112678580815
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3072,8192,0.2710079881880018
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,16384,32,0.0031715554909573663
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,2560,5120,0.09827733039855957
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2561,256,65536,0.3759893311394586
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,512,0.011583111352390714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,768,6144,0.012005333271291522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,4096,128,0.021352888809310064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,16384,3072,0.2483688990275065
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,128,512,0.009195555415418413
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,7168,16384,0.06851289007398817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,768,1024,0.006169777777459886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,51200,65536,0.9692853291829427
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,16384,51200,0.4273306793636746
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,3584,3584,0.12250755892859565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1024,16384,4096,0.09800889094670613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3,512,2560,0.007024889190991719
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,65536,51200,0.9508800506591797
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,51200,5120,0.1547573275036282
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,2048,256,0.00328177772462368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,33,2048,12288,0.020999110407299463
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2048,12288,0.2287893295288086
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,768,16384,0.10891111029518975
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,3584,3072,0.041998220814598926
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,8192,51200,1.8957733578152125
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,128,1536,0.014843554960356818
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,1024,2048,0.007087111473083496
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,6,32,3072,0.005919999960396025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3072,256,0.005408000200986862
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,49,6144,256,0.0034426665968365143
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,12288,12288,0.06018400192260742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,128,65536,0.06775733497407702
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2048,8192,0.019004444281260174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,5120,5120,0.07767199807696872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,1536,768,0.006220444622966979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,32,0.006719111154476802
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,3584,3584,0.012449777788586087
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,10240,4096,2.288275612725152
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,5120,6144,0.0343484448062049
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,512,7168,0.03779822256830003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,3584,0.0914728906419542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3584,128,0.00397511116332478
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,3072,8192,0.1560373306274414
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,10240,768,0.008480889101823172
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,2048,3072,0.009273777405420939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,65536,2560,0.11160355144076878
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,32,51200,0.08503466844558716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,1536,128,0.009768889182143742
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,6144,1024,0.0064026667839950984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,64,2048,0.019331556227472093
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,16384,5120,64,0.05369866556591458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,128,4096,0.008575999902354347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,2048,7168,0.01889333294497596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,1024,5120,0.07876355780495538
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,5120,2048,0.0843377775616116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,256,51200,0.06177155839072334
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,160,3072,512,0.00574933323595259
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,4096,1024,0.008691555923885768
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,32,2048,0.006053333481152852
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,4096,51200,0.11511466238233779
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,32,0.0037208890749348533
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,2048,8192,0.14878666400909424
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3584,128,0.006168889088763132
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3073,4096,256,0.017671111557218764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,129,51200,1536,0.07156533002853394
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,3584,2048,0.007348444726732042
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,8192,16384,0.9005973603990344
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2560,8192,1536,0.06327555576960246
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,5120,3072,0.11911022663116455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,256,1024,0.011976888610257043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3584,3072,0.07148000266816881
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,2560,4096,0.026403556267420452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,128,768,0.0058355554938316345
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,1536,12288,0.028175112273957994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,512,3072,0.020434666011068556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,4096,2048,0.06423733631769817
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4097,256,8192,0.10794488588968913
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,1024,32,0.0029111111329661477
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,1536,16384,0.03431288732422723
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,10240,6144,0.6854151089986166
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,128,6144,0.050775110721588135
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,128,256,51200,0.01770755483044518
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,193,3584,12288,0.05173066589567396
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,1536,5120,0.012463111016485425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,65536,3584,0.33346133761935765
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,161,5120,5120,0.02623733381430308
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,8192,32,0.002936888900068071
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,7168,2560,0.12225866317749023
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,51200,51200,0.7660266558329264
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2560,7168,16384,0.5726533465915257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,65536,10240,0.20316267013549805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,512,4096,0.06938933663898043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,113,5120,7168,0.024950222836600408
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,6144,12288,0.3544791009691026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,512,512,0.003359111025929451
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,6144,51200,0.09334399965074326
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,5,4096,512,0.00564444437623024
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,768,1536,0.006719999843173557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,2560,8192,0.12211644649505615
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,112,2048,51200,0.059092442194620766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3328,1024,51200,0.4992684258355035
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,10,8192,16384,0.07488266626993816
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,113,5120,128,0.006179555422729916
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,6144,64,0.00638755535085996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,51200,8192,2.7019013298882375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,769,1024,2560,0.01293333371480306
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1024,512,0.009217778013812171
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,256,8192,0.022504000200165644
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,3584,0.021329777108298406
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4097,2048,32,0.008375999828179678
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,14,2560,256,0.004446222136418025
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,32,51200,0.007301333049933116
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,3584,16384,0.6648586591084799
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,1536,768,0.11861333582136367
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,256,3584,0.041830221811930336
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,193,51200,128,0.017471111483044095
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,512,0.005092444519201915
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2049,12288,12288,0.815799978044298
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,3584,3072,0.011072888970375061
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,10240,51200,0.16175556182861328
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,768,4096,64,0.004497777670621872
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,16384,0.8698719872368706
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,12288,10240,1.4354933632744684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4353,65536,5120,4.40757793850369
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,512,10240,0.0362737774848938
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,6144,2560,0.01427111029624939
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,81,1536,64,0.0069288888739215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3329,2048,8192,0.1746915578842163
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,256,7168,0.031714667876561485
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,5120,2048,0.010792888700962067
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,160,51200,3584,0.11715377701653375
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,3584,1536,0.009279111193286048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,128,128,0.007984889050324758
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,224,7168,1024,0.011609777808189392
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,4096,256,0.00534400012758043
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2305,7168,768,0.04907200071546766
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,51200,32,0.004273777620659934
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8,8192,8192,0.027735999888843958
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,3072,10240,0.020939555433061387
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,768,0.007027555671003129
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,2048,3584,0.012624888784355588
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,8192,7168,0.029736889733208552
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,1536,7168,0.01616533266173469
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,4096,1536,0.029641777276992798
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,6144,8192,0.24739289283752441
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,256,7168,0.020061333974202473
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,768,32,0.0027075554761621687
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,113,2048,1536,0.006510222123728857
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,11,6144,3072,0.012964444855848948
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,3584,3.1880213419596353
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,3072,3072,0.01608711149957445
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,51200,1536,0.40449955728318954
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,65,12288,7168,0.03573866685231527
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,65536,512,0.015419556034935845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,112,1536,8192,0.012934221989578672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,5120,3584,0.07016355461544461
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,10240,512,0.053301334381103516
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4,7168,65536,0.2472017871008979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,48,3584,8192,0.016899555921554565
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1793,16384,512,0.040335112147861056
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,2048,512,0.017981333865059745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,32,512,0.0034657776769664553
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1025,3072,7168,0.06545155578189425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3328,6144,51200,1.944943110148112
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,256,8192,0.029648890097935993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1024,256,0.007898666792445714
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,256,2560,0.06456622150209215
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,65536,1.3210977978176541
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,112,128,2048,0.0063822223908371395
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,3584,32,0.0027804445061418745
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,769,16384,1024,0.03784977727466159
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,80,128,512,0.005541333307822545
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,64,0.0063564446237352155
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,3584,32,0.012401777837011548
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1,10240,1024,0.010607110957304636
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,2560,65536,1.4682444466484919
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,1536,4096,0.05998400184843275
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,12288,0.8765698009067111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16,3584,12288,0.02494755552874671
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,96,1024,32,0.006314666734801398
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,13,6144,2560,0.01218933363755544
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,1536,51200,0.2942906750573052
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,65536,32,0.004867555366622077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,10240,1536,0.09369066688749526
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,768,65536,0.042240000433391996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,12288,2560,0.01888088881969452
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,51200,3584,0.09835733307732476
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,7168,8192,0.23771821127997506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2561,12288,6144,0.5347173478868272
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,6144,8192,0.02959200077586704
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,16384,7168,0.0565626687473721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,2560,12288,0.025118221839269
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32768,7168,1536,0.591268433464898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,161,1024,32,0.006809777683681912
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1280,12288,64,0.015678221980730694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,1024,10240,0.1150897741317749
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,768,1536,10240,0.04184622234768338
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,3072,1024,0.004941333499219683
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,512,12288,0.016701334052615695
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,2048,7168,0.06257422102822198
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,16384,10240,0.057068447271982826
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,65,2560,65536,0.0914542211426629
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,4096,256,0.0031973332580592898
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,32,32,0.007195555501514011
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,5,10240,4096,0.027190221680535212
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,7168,64,0.0028400001012616684
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2817,10240,65536,3.921522776285807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,16384,1536,0.015561777684423657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,16384,16384,10240,4.726001739501953
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,80,512,8192,0.01994222237004174
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,64,8192,0.01870133313867781
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,17,10240,1024,0.00981422182586458
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,2048,768,0.009200889203283522
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,64,6144,0.013567110730542077
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1792,12288,51200,3.526709238688151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,8192,768,0.010007999837398529
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,192,3072,3584,0.01850666602452596
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,9,6144,6144,0.0247688889503479
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,6144,768,0.04990933338801066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4096,16384,12288,1.269984033372667
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,14,8192,6144,0.02488888965712653
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3072,1536,3584,0.08219822247823079
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3584,3072,7168,0.16190044085184732
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2,2560,8192,0.016069332758585613
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,1024,2560,0.03222222129503886
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,97,5120,6144,0.025803556044896443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4352,3072,10240,0.2720853222741021
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,96,128,512,0.004699555536111196
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,10240,0.14204711384243435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,768,8192,0.016536000702116225
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,224,6144,3072,0.01831822262869941
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,32,65536,1024,0.025478222303920325
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,512,3584,0.05469510952631632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,16384,128,0.004961777892377642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,1024,2048,0.010743111371994019
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,10240,64,0.01331822243001726
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,768,16384,0.021903110875023737
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,128,6144,0.010269333091047075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,7168,51200,2.5952559577094183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,3072,3584,0.01088711122671763
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,1024,0.0369395547442966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,1536,2560,0.028152889675564233
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,7168,1.3338346481323242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,4096,7168,0.10944088962343003
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,1024,2560,0.005985777825117111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,225,32,1024,0.003903999924659729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,97,64,51200,0.0544471111562517
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,12,64,51200,0.008533333738644918
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,5120,64,0.011237333218256632
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,128,51200,6144,0.09691022502051459
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,6,64,65536,0.06271644433339436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,48,4096,128,0.002954666606254048
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2304,8192,2560,0.09385955333709717
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,161,12288,65536,0.4484693209330241
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,6144,2048,0.14168266455332437
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,224,3072,12288,0.038619554705089994
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,1024,12288,0.062126219272613525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1024,7168,8192,0.17309599452548555
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,32,512,0.008971555365456475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4097,6144,64,0.02980888883272807
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,512,0.22591911421881783
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1,768,2560,0.0069413334131240845
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,1024,512,0.0046364445653226645
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2049,51200,128,0.10160888565911187
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4352,256,6144,0.08972266647550796
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,3584,7168,0.01587288909488254
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8,51200,1024,0.021516443954573736
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,4096,768,0.006107555495368109
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1537,128,1536,0.01292622254954444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,4096,6144,0.023594665858480666
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,5120,7168,0.022292445103327434
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2817,1536,768,0.012891555825869242
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1793,3584,256,0.00909511082702213
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,65536,6144,0.11032711135016547
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3073,1024,5120,0.09356711308161418
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,7168,10240,0.051078223519855075
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,32,12288,0.006187555690606435
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,16,3584,3072,0.011064888702498542
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2816,4096,7168,0.17616355419158936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4096,256,7168,0.02785244584083557
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,12,10240,256,0.0064533332155810455
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,9,2048,2048,0.010809777511490716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,5120,5120,0.017811555001470778
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1537,2048,2048,0.02604533400800493
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,769,12288,65536,1.2740507125854492
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,512,16384,3584,0.08816888597276475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2304,16384,3072,0.17120888498094347
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,3072,1536,0.012279999752839407
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,2048,8192,0.13986666997273764
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,257,65536,1024,0.066721777121226
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,13,65536,128,0.007363555332024892
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,129,3072,3072,0.01648177703221639
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1536,1536,512,0.01108000013563368
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,256,3584,1024,0.006362666686375936
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,1536,64,0.005991111199061076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,257,128,6144,0.013760888742076026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2048,6144,2560,0.0889973309304979
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3841,1536,12288,0.18986933761172822
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,11,3584,2048,0.008243555823961893
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,5120,2560,0.15955644183688694
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,3072,0.0646373364660475
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2560,4096,64,0.008531555533409119
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,33,4096,5120,0.013063111239009433
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,10,6144,2560,0.011701333026091257
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,15,5120,5120,0.014329777823554145
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,14,6144,512,0.0037911112109820047
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1280,12288,7168,0.1856844425201416
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,49,65536,8192,0.1466053326924642
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,7,1536,65536,0.05539733171463013
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,256,6144,3584,0.018831110662884183
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3329,32,2560,0.021607110897699993
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,65,3584,3584,0.013038222160604266
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,128,16384,3584,0.0236453331179089
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4,51200,2560,0.044970668024486966
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,4352,12288,1024,0.15967733330196804
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1537,5120,2560,0.057387557294633657
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,32,16384,0.018648000227080453
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,513,65536,2048,0.20499466525183785
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3841,2560,16384,0.3139502207438151
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,512,3584,51200,0.17016000217861602
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,256,3584,7168,0.02219911085234748
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,128,6144,0.049920890066358775
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3585,512,5120,0.04334577918052673
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3585,7168,12288,0.8640809059143066
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1,7168,1024,0.007763555480374231
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,1536,4096,0.05570755402247111
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3072,128,256,0.008858666651778752
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1536,1536,2560,0.028740445772806805
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,256,128,0.005894222193294101
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1281,768,256,0.006440889090299606
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,2560,4096,0.012689777546458773
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,65536,256,0.011689778003427716
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2,256,4096,0.008112000094519721
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2817,6144,16384,0.5954631169637045
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,1024,8192,0.036976890431510076
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32,2560,3584,0.010240000155236986
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,129,16384,128,0.004600888739029567
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,12288,2048,0.10210755798551772
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,2048,6144,1.1076631546020508
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,192,7168,16384,0.07426755295859443
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,7,6144,5120,0.01847288840346866
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,12,5120,1536,0.009670221971140968
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,64,6144,65536,0.23271467950608996
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,6,768,12288,0.022273777259720698
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,49,128,51200,0.05346399876806471
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,512,1024,3072,0.013648889131016202
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,3072,1024,0.05118844575352139
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,17,512,8192,0.01183733344078064
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,96,3072,8192,0.02093066606256697
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,2304,32,256,0.0030008889734745026
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3840,32,128,0.0069982219073507525
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2048,6144,7168,0.1339084439807468
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,8192,256,5120,0.07304977708392672
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,192,3072,65536,0.1291102170944214
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,32,1024,65536,0.07653688722186618
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3840,6144,3584,0.16055822372436523
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,64,6144,128,0.004644444419278039
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,16384,2048,0.018571555614471436
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,10,32,6144,0.010063110954231685
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,1536,0.03945066531499227
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,4,1024,6144,0.010452444354693094
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,3073,5120,3072,0.0952755543920729
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,4096,7168,0.14495378070407444
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,1793,3072,65536,0.8255697886149088
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,1536,5120,768,0.021337777376174927
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3,10240,4096,0.02777333392037286
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,4353,51200,12288,4.6111776563856335
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,7168,128,0.009861333502663506
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,3584,1536,768,0.014375110467274984
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,2048,0.18841244114769828
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,13,5120,2560,0.010404444403118556
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8,2305,16384,65536,6.019417656792535
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,768,1536,5120,0.030750221676296655
SGLang,0.5.10,NVIDIA H200,gemm,sglang,float16,32768,51200,10240,53.636823866102425
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,2049,2560,6144,0.09666844209035237
SGLang,0.5.10,NVIDIA H200,gemm,sglang,fp8_block,3585,256,65536,0.802278200785319
